Potential in Systematically Placing a High Capacity of PV and
Wind Sources in Germany
Storage and Transmission Capacity Requirements Analysis and a Case
Study on Morocco
BY
Arabi Sadeq Mohammad Abdelhaq
Submitted to
The Faculty of Engineering in Cairo University and Faculty of Electrical
Engineering/Computer Science in the University of Kassel in partial fulfillment of the
requirements for the degree of Master of Science in Renewable Energy and Energy
Efficiency for the Middle East and North Africa Region
Kassel, February 2012
Supervised By
Dr. Stefan Bofinger
Fraunhofer - IWES
M.Sc Raphael Spiekermann
Fraunhofer - IWES
Reviewed By
Prof. Dirk Dahlhaus
University of Kassel
Prof. Mohammed Salah ElSobki
Cairo University
DISCLAIMER
To the best of my knowledge I do hereby declare that this thesis is my own work. It has not been
submitted in any form of another degree or diploma to any other university or other institution of
education. Information derived from the published or unpublished work of others has been
acknowledged in the text and a list of references is given.
Kassel,
28.02.2012,
Arabi Abdelhaq
ii
Executive Summary
The world is facing new challenges related to depleting natural resources, expanding economies,
and an ever-growing population with resource-intensive life styles. Accordingly, policymakers
must actively analyze different combinations of available renewable energy technologies as
viable solutions to these challenges. The ultimate objective is to provide diverse options for the
effective utilization of renewable energy both in the present and the near future.
This MATLAB-based technical study analyzes the potential of a systematic-optimized placement
of wind turbines and Photovoltaic panels (PV) to minimize the requirements for the energy
supply structure by reducing the residual load in Germany. A simpler case study is also presented
for Morocco.
This study seeks to analyze the installed capacity required to ensure that PV panels and wind
turbines contribute to reduce significantly the residual load in Germany based on the allocation
methods used in this study. In addition, Germany was divided into six regions to see the impacts
of optimizing the installation of PV and wind capacity in small regions to the whole system,
allowing us to take into consideration transport and storage capacity requirements needed to
consistently fulfill the maximum load demand.
The “six regions” scenario proposed in this study, with a total installed capacity of 238.01 GW
reduces the average residual load by 69.18%. This is an improved figure from the 65% residual
load reduction by the other Germany as one region scenario. Moreover, load flow with a mean
value of 1,474 MW for the lines transmission capacity is established through a Direct Current
(DC) grid between the regions to balance all the hourly lack and excess of power before using
storage of a 17 GW range and a capacity of 141.737 TWh to completely make up for the hourly
differences between load demand and generation.
These results for storage capacity requirements and transport were compared to two scenarios:
Germany as one region scenario and another reference scenario, and results showed that
maximum transport capacity needed is 27% and 12% less than the two scenarios respectively and
capacity storage requirements are 13% and 2% less respectively. This demonstrated the
usefulness of using the smaller regions model as an optimization approach to reduce storage and
transmission requirements taking into account the whole supply structure.
The case study on Morocco provides a theoretical estimation of the installed PV and wind
capacity required to fully cover the yearly energy consumption and was found to be 19,520 MW
and 12,250 MW for PV and wind respectively. Storage requirements are studied for each case
and show a clear differentiation between the characteristics of this storage for each source were
PV accumulated storage requirements are 27% less than wind.
iii
Acknowledgments
This study would not have been possible without the support of many people. I would like to
express my gratitude to my supervisors in Fraunhofer institute in Kassel, Dr. Stefan Bofinger for
the opportunity to work on this project, and Raphael Spiekermann who was abundantly helpful
and offered invaluable assistance, support and guidance whenever needed.
I would also like to thank my supervisors in the University of Kassel and Cairo University, Prof.
Dirk Dahlhaus and Prof. Mohammed ElSobki for their feedback and insight.
Thanks are also conveyed to the REMENA program and DAAD for providing the administrative
and financial means for me to take part in this masters program and accomplish this thesis work.
Special thanks also to all my friends both in the REMENA masters program and Fraunhofer
institute for their unconditional support throughout the course of the master thesis work and also
to the people who have helped me in the editing phase of my thesis from whom I have learnt
a lot.
I wish to express my love and gratitude to my beloved family, for their understanding and
endless love, through the duration of my studies in Germany, their words and encouragement
have been always a “renewable” source of comfort and support.
Finally, I would like to more than thank my friend Rand Al‟Zubi for everything she supported
me with, whether it was words, ideas or giving me the extra push when needed. You were not a
best friend, but the best kind of a friend. Shokran.
iv
Contents
DISCLAIMER ................................................................................................................................ ii
Executive Summary...................................................................................................................... iii
Acknowledgments.......................................................................................................................... iv
List of Figures ............................................................................................................................... vii
List of Tables ................................................................................................................................. ix
Abbreviations and Acronyms ......................................................................................................... x
1.
Introduction ............................................................................................................................. 1
1.1.
Motivation of the work ..................................................................................................... 1
1.2.
Problem Statement and Research Objectives ................................................................... 1
1.3.
Hypothesis ........................................................................................................................ 2
2.
A Brief Literature Review ....................................................................................................... 1
3.
Technical Background ............................................................................................................. 4
3.1. Historical Background on the available PV and Wind energy sources situation in
Germany and their complementarities ........................................................................................ 4
3.2.
4.
3.2.1.
The Correlation Coefficient ...................................................................................... 5
3.2.2.
Covering the Residual Load...................................................................................... 6
3.2.3.
Covering the Residual Load Peaks ........................................................................... 7
3.2.4.
The Suitability Factor ............................................................................................... 8
The Model.............................................................................................................................. 10
4.1.
Input Data................................................................................................................ 10
4.1.2.
Wind, Solar and Load Demand Data ...................................................................... 11
The Program Algorithm ................................................................................................. 15
Analysis of Germany as One Region .................................................................................... 20
5.1.
Simulation Results.......................................................................................................... 21
5.1.1.
PV Capacity Installed ............................................................................................. 22
5.1.2.
PV Capacity Installed Tilt Angles and Orientation Preferences ............................. 23
5.1.3.
Wind Capacity Installed .......................................................................................... 25
5.2.
6.
Data Used ....................................................................................................................... 10
4.1.1.
4.2.
5.
PV and Wind Energy Allocation Methods ....................................................................... 5
Discussion and Conclusion of the results ....................................................................... 26
Analysis of Germany Divided into Regions .......................................................................... 30
v
6.1.
Simulation model ........................................................................................................... 33
6.1.1.
Experimental framework ........................................................................................ 33
6.1.2.
Simulation results.................................................................................................... 34
6.1.3.
6.2.
7.
8.
Discussions and Conclusion of the Results of the Simulation Model ........................ 45
The Direct Current Load Flow Matrix – Transport Capacity ........................................ 46
6.2.1.
The DCLFM and Load Flow Parameters ................................................................... 51
6.2.2.
Duration Curves of the Transmission Performance (DCTP)...................................... 52
6.2.3.
Discussions and Conclusions of the Results of the Transport Capacity..................... 56
Comparison with the Reference Scenario and Conclusions .................................................. 57
7.1.
Regional Scenario versus Reference Scenario ............................................................... 57
7.2.
Regional Scenario versus Germany as One Region Scenario ........................................ 59
7.3.
Conclusions of the Comparison ..................................................................................... 61
Case study: Morocco ............................................................................................................. 63
8.1.
Data ................................................................................................................................ 64
8.1.1.
Solar Data................................................................................................................ 64
8.1.2.
Wind data [14]........................................................................................................... 64
8.1.3.
The load curve......................................................................................................... 65
8.2.
8.3.
Experimental Framework ........................................................................................... 66
Simulation ...................................................................................................................... 67
8.3.1.
Wind Scenarios ........................................................................................................... 67
8.3.2.
PV scenarios ............................................................................................................... 68
8.3.3.
Storage Requirements ................................................................................................. 72
Discussion of the results and conclusions ................................................................................. 74
9.
Conclusions and Recommendations for Further Research .................................................... 75
9.1.
Conclusions .................................................................................................................... 75
9.2.
Recommendations for Further Research ........................................................................ 77
References ..................................................................................................................................... 78
Appendices .................................................................................................................................... 80
Appendix A: DC Load Flow ..................................................................................................... 80
Appendix B: Central storage model for the six regions ............................................................ 81
Appendix C: Installed Wind and PV Capacity for Regions ...................................................... 82
vi
List of Figures
Figure 1: Map of Germany Divided into 3027 Pixels..................................................................... 2
Figure 2: Residual load for Germany with 10% over Maximum Residual Load Limit ................. 7
Figure 3: PV vs. Wind Generation in Pixel One ........................................................................... 11
Figure 4: PV Potential According to the CORINE Land Cover Data .......................................... 13
Figure 5: PV and Wind Potential in Germany .............................................................................. 14
Figure 6: Load of Germany in 2009 ............................................................................................. 15
Figure 7: Main Inputs and Assumptions Used in the Model ........................................................ 18
Figure 8: Model Steps to Simulate the Results ............................................................................. 19
Figure 9: RLDC for Germany ....................................................................................................... 22
Figure 10: PV Capacity Installed for Germany ............................................................................ 23
Figure 11: Tilt and Orientation Angles Used to Allocate PV Panels ........................................... 24
Figure 12: Frequency of Tilt and Orientation Angles Use of the Different Installed PV Capacity
....................................................................................................................................................... 24
Figure 13: The Installed Wind Capacity ....................................................................................... 25
Figure 14: Half the Installed Wind Capacity for Germany........................................................... 26
Figure 15: Installation of PV and Wind Capacity ......................................................................... 27
Figure 16: Different Time Ranges for Residual Load versus Generation Capacity for Germany 28
Figure 17: Allocation Methods Change for Germany and Suitability Factor Decrease in
Analyzing These Different Results ............................................................................................... 29
Figure 18 the Six Regions with Population Densities in Colored Code ....................................... 32
Figure 19: Installed PV Capacity in the South-West Region ....................................................... 36
Figure 20: Orientation and Tilt angles for the Installed Capacity of Region One (South-West) . 36
Figure 21: Comparison between Frequency of Tilt and Orientation Angles used for the Different
Installed PV capacity in Region one (South-West) ...................................................................... 37
Figure 22: Installed PV Capacity in the Mid-West Region .......................................................... 38
Figure 23: Orientation and Tilt angles for the Installed Capacity of Region Two (Mid-West) ... 38
Figure 24: Comparison between Frequency of Tilt and Orientation Angles used for the Different
Installed PV Capacity in Region One (Mid-West) ....................................................................... 39
Figure 25: Installed PV Capacity in the South-East Region ......................................................... 39
Figure 26: Orientation and Tilt Angles for the Installed Capacity of Region Four ...................... 40
Figure 27: Comparison between Frequency of Tilt and Orientation Angles used for the Different
Installed PV Capacity in Region Four (South-East) ..................................................................... 40
Figure 28: Installed Wind Capacity in the North-West Region .................................................... 41
Figure 29: Installed Wind Capacity in the Mid-East Region ........................................................ 42
Figure 30: Installed Wind Capacity in the North-East Region ..................................................... 42
Figure 31: Allocation Methods Change for Region One and Suitability Factor Decrease in
Analyzing the Different Results.................................................................................................... 44
Figure 32: Installation Order for PV and Wind Shares in the South-West Region (Region One) 45
Figure 33: Installation Order for PV and Wind Shares in the North-East Region (Region Six) .. 45
vii
Figure 34: The DC Network between the Proposed Six Regions Shown According to their
Geographical Orientation .............................................................................................................. 47
Figure 35: Residual Load Duration Curve for the Regional Storage............................................ 49
Figure 36: Hour Percentage where there is a Positive Residual Load per Region ....................... 49
Figure 37: Hourly Residual Loads for One Year, with Positive Residuals Above the Red Line
and Negative Residuals Below it .................................................................................................. 50
Figure 38: DCTP for Line 1 with Absolute Values ...................................................................... 53
Figure 39: Histogram for the Frequency of Use of Line 1 Capacity ............................................ 53
Figure 40: DC Load Flow in Germany with Bypass Storage ....................................................... 55
Figure 41: PV Generation Difference between the Installed Regional Capacity and the Reference
Scenario......................................................................................................................................... 57
Figure 42: Selected Points for PV (Black-Dispersed) and Wind (Green-Concentrated).............. 63
Figure 43: Power Curves at a Specified Location in Morocco ..................................................... 64
Figure 44: The Load Curve of Germany and Morocco for 2009 and 2010 Respectively ............ 65
Figure 45: Installed Wind Capacity for Morocco ......................................................................... 67
Figure 46: 100% Wind Share Scenario for the Coverage of the Residual Load of Morocco ....... 68
Figure 47: No PV Generation at Night time and Comparison with the Load............................... 69
Figure 48: Installed PV Capacity of 19520 that is Able to Bring the Mean Residual Load to Zero
....................................................................................................................................................... 70
Figure 49: Tilt and Orientation angles for the PV scenario .......................................................... 71
Figure 50: Frequency of Tilt and Orientation Angles for the PV Scenario .................................. 71
Figure 51: PV Generation Curve for a 100% PV Scenario .......................................................... 72
Figure 52: Hourly Theoretical Accumulated Storage for PV and Wind ...................................... 72
Figure 53: Storage Requirements for 100% PV and Wind Scenarios in Morocco ....................... 73
Figure 54: Power flow from X to Y through Z ............................................................................. 80
Figure 55: RLDC for line 1 (Connecting Centers of Regions 1-2)............................................... 81
Figure 56: Installed PV and Wind Capacity for All regions where Source is not Leading in
Installed Capacity.......................................................................................................................... 82
viii
List of Tables
Table 1: Terms Used in the Suitability Factor Calculation for PV and Wind ................................ 9
Table 2 Academic Division of the Regions in Germany .............................................................. 10
Table 3: Assumptions and Pre-Set Values Used in the Model ..................................................... 16
Table 4: Percentages of Load Coverage in Germany ................................................................... 20
Table 5: Simulation Results for Germany as One Region ............................................................ 21
Table 6: Percentages of Load Coverage in Germany by the Six Newly Established Regions ..... 31
Table 7 Average Residual Loads and the Leading Source in Each Region after Installing Half the
Available Capacity ........................................................................................................................ 34
Table 8: Centrally Located Storage Scenario and DC Lines Distances (km) ............................... 48
Table 9: DCLFM Matrix for the Six Regions for a Distant Storage Configuration ..................... 51
Table 10: LF Indicators for each Line .......................................................................................... 52
Table 11: Optimization of Load Flow Capacity Requirements .................................................... 54
Table 12: Differences in Installed PV Capacity between the Reference Scenario and Simulation
Results for the Six Regions ........................................................................................................... 58
Table 13: Differences in Installed PV Capacity between the Reference and Six Regions
Scenarios ....................................................................................................................................... 59
Table 14: Differences in installed PV Capacity between Germany as One Region Scenario and
Simulation Results for the Six Regions ........................................................................................ 59
Table 15: Differences in Installed PV Capacity between Germany as One Region Scenario and
Simulation Results for Six Regions .............................................................................................. 60
Table 16: Summary of Main Conclusions between the Regional Scenario and the Reference
Scenario as well as Between the Regional Scenario and Germany as One Region Scenario ...... 62
Table 17: Load Demand Characteristics in Morocco and Germany............................................. 66
Table 18: Scenarios for the Coverage of the Residual Load of Morocco by PV and Wind ......... 66
Table 19: Summary of Wind Scenario Results ............................................................................. 67
Table 20: Comparison between the Two PV Results Based on Different Weighing Factors ...... 69
Table 21: Line Parameters for the Central Storage Model ........................................................... 81
ix
Abbreviations and Acronyms
CC: Correlation Coefficient
DC: Direct Current
DCFLM: Direct Current Flow Load Matrix
DCTP: Duration Curves of the Transmission Performance
GHI: Global Horizontal Irradiance
LF: Load Flow
MENA: Middle East and North Africa
PTDF: Power Transfer Distribution Factor
PV: Photovoltaic
RES: Renewable Energy Sources
RL: Residual Load
RLDC: Residual Load Duration Curves
TP: Transmission Performance
x
1. Introduction
1. Introduction
1.1. Motivation of the work
Germany aims to source 80% of its electrical demand from renewable sources by 20501 and
studies show that a 100% renewable energy share is even possible by then2. Accordingly,
analyzing the transformation of Germany‟s energy supply system by fluctuating renewable
energy serves as a valuable model of shifting towards a highly renewable energy scenario.
In order to reach a high renewable energy share of the total energy generation in Germany, one
needs to consider the energy consumption, study the available resources, and employ sound
methodologies required for putting renewable energy plants in place.
The primary challenge in this analysis is to match Germany‟s energy demand to PV-wind energy
generation to accommodate supply excess and shortages. To overcome this problem, large
storage and transmission capacity will be needed. We must keep in mind, that the allocation of
PV and wind must not be based only on where the sources are available as it is not necessarily
the economic optimum.
Therefore, sound allocation methods of these two sources must be studied to consider for other
factors: the percentages of the unused excess energy, the contribution to load covering, and
fitting the generation to the load and reducing the load peaks.
In Morocco, with its low electrical demand, and its high solar radiation and wind resources, the
focus was to get an idea of its export capacity and storage requirements from each resource in a
100% renewable scenario.
1.2. Problem Statement and Research Objectives
The research objectives are to make qualitative predictions that answer these questions:
Could we improve meeting the load demand by specific site selection?
Could we significantly reduce waste due to excess energy generation by specific site
selection and equalize yield losses caused by the selection?
Could we reduce the demand for transmission and storage capacity?
What effects on these qualitative predictions could we get if we move from a Germany as
one region scenario to a regional scenario to make an “optimized” site selection?
1
Press release No. 012/11 by the Federal Ministry for the Environment, Nature Conservation and Nuclear Safety
(BMU)
2
Study by Fraunhofer Institute: 100% renewable electricity supply by 2050
1
1. Introduction
By answering these questions, decision-makers in Germany and Morocco shall be able to decide
on new incentives for future allocation of major wind-PV farms, and should be able to develop
indicators to raise awareness about the limitations and possibilities to reach the 100% renewable
energy using PV and wind. In addition, policy makers should have an indication about the
required infrastructure for the new Direct Current (DC) network to cover the load based on the
results of this study.
It is important to note, that this study is a first approach for an improved site selection, only three
factors for optimization are considered which will be discussed in detail, and there should be far
more to get a better simulation quality. Therefore, the results are only qualitative hints to the
effect of a specific site selection.
1.3. Hypothesis
It is possible that wind and PV sources can address a large part of the energy consumption needs
in Germany and reduce the residual load by following specific - well managed - energy
allocation methods that take into consideration the load demand profile and its matching with the
available PV-wind share. Furthermore, it is possible to optimize this matching and reduction in
residual load by analyzing on a micro-level, different regions within Germany and including
transfer and storage possibilities.
Figure 1 shows Germany divided into the small pixels used in this study to simulate different
results3
55
54
53
Latitude
52
51
50
49
48
47
4
6
8
10
12
14
16
Longitude
Figure 1: Map of Germany Divided into 3027 Pixels
3
A definition of the “pixel” will follow in chapter 4
2
2. A Brief Literature Review
2. A Brief Literature Review
Several studies have researched a 100% renewable mix between various renewable energy
resources. The main difference between these studies has been the geographical location of the
study, and the type of renewable sources chosen to investigate while applying different research
methodologies.
In the study of Joakim Widén[1] on the correlations between large-scale solar and wind power in a
future scenario for Sweden, the author studied the correlation between wind and PV farms in
different parts of the country.
The smoothing effect was thus analyzed by using the sample correlation between wind and solar
power as an allocation strategy for wind and PV farms.
The time variability and coincidence in certain locations proved that there is a negative
correlation between solar and wind outputs and that using large-time scales provides for the
strongest correlation
Another study by Dominik Heide, Lueder von Bremen, Martin Greiner, Clemens Hoffmann,
Markus Speckmann, and Stefan Bofinger[2] presented results similar to the research presented in
this study, although it focuses on the EU. The results drawn highlight the optimal mix between
PV and wind in a highly renewable scenario standing at 55% wind, 45% PV shares and examine
the effect on the storage size in such a mixed share scenario as opposed to a wind-only or PVonly scenario.
A study by Thomas Nikolakakis and Vasilis Fthenakis [3] focused on New York State, and
examined the New York state grid and the maximum penetration possibilities by different RES,
taking into consideration different scenarios and mix shares.
In his study conducted in Spain, Ghassan Zubi [4] concluded that at least 80% PV and Wind share
proved to be the basis for any economical highly renewable futuristic scenario, and a technical
mix of 30% Wind and 15% PV.
3
3. Technical Background
3. Technical Background
3.1. Historical Background on the available PV and Wind energy sources
situation in Germany and their complementarities4
Germany has a leading role in the development and employment of PV and wind renewable
energy sources. It topped the rankings of countries based on the total installed wind capacity by
the end of 2011 with 28.576 GW5 installed wind capacity, meeting around 6.3%6 of its load
demand in addition to 24.8 GW7 of installed PV capacity by the end of 2011.
In 2010, Germany added more PV capacity than all the world capacity installed in 2009.With the
additional investment in renewable energy, Germany ranked first for PV and fifth for wind
technology capacity added annually. As for its existing capacity, Germany ranked first for PV
and third for wind.
Germany met 11% of its final energy consumption with renewable resources in 2010. Wind
power accounted for nearly 36% of renewable generation, followed by biomass, hydropower,
and solar photovoltaic (PV).[5]
In Germany, as anywhere else, the issue of complementarities between PV and wind power
generation is important; the better correlated PV and wind generation are with the load, the more
beneficial it will be for the grid to accommodate this share. Moreover, the generation of energy
based on these resources is directly dependent on weather conditions. On the other hand, the
spatial distribution of these resources has an effect on the correlation with the load.
PV and wind resources are intermittent resources; there is no PV generation at night, and its
generation increases and decreases daily, depending on weather conditions, cloud formation, and
shading, among other factors. Alternatively, wind generation is less predictable within a certain
time period compared to PV and has sharper generation curves than PV due to its high temporal
dependency.
A higher correlation between wind and PV with load demand reduces storage requirements of
the generated energy. Therefore, this study focuses on the effect of correlating PV and wind
sources with the load.
4
The figures in the second and third paragraphs are taken from the reference mentioned at the end of paragraph 3
Source: http://windmonitor.iwes.fraunhofer.de/
6
This percentage is extrapolated from 6% load met found in reference 5 for the year 2010.
7
Press Release: Zubau an Photovoltaik-Anlagen 2011 noch höher als im Rekordjahr 2010 http://www.bundesnetzagentur.de. 7.5 GW from this reference were added to the figure from reference 5 to make
2011 figure.
5
4
3. Technical Background
3.2. PV and Wind Energy Allocation Methods
Three methods are applied to determine the allocation of wind turbines and PV modules: The
correlation coefficient method, covering the residual load method, and covering the residual load
peaks method.
A „suitability factor‟, described later in the section, is the formula that connects and weighs the
contribution of these three methods for the allocation of different PV and Wind capacity.
3.2.1. The Correlation Coefficient
The correlation coefficient (ρ) between two data sets quantifies the relationship between these [6],
and can be seen mathematically from the following formulas, as the covariance of the two data
sets we have, divided by the product of their standard deviations.
In this study, the two data sets are the PV power generation time-series, (pv), against the load
demand time-series, and second, between the wind power generation time-series (w) against the
same load demand time-series.
Equations 1 and 2 are the mathematical forms of the correlation coefficient, where the numerator
describes the covariance of the data, Equations 3 and 4, and the denominator describes the
variance of the two data sets, namely PV against load demand and wind against load demand.
( pv, l )
( w, l )
C ( pv, l )
C ( pv, pv)C (l , l )
1
C ( w, l )
C ( w, w)C (l , l )
2
Where,
T 8760
C ( pv, l )
( pvt
t 1
pv
)(lt
l
)
3
5
3. Technical Background
T 8760
C ( w, l )
( wt
w
)(lt
l
)
4
t 1
In Equations 5, 6 and 7, µpv, µw and µl are the mean values of the PV, wind, and load demand
data sets respectively.
1 T 8760
( pvt )
T t1
pv
1 T 8760
( wt )
T t1
w
1 T 8760
(lt )
T t1
l
5
6
7
This method is used to indicate us how well the wind or PV power generation time-series
correlate (or are in harmony) with the load demand time-series and is measured with a value
ranging from of 1 to -1, whereby1 means that the variables are totally correlated and -1 means
that variables are negatively correlated between these two sets. Equation 8 states:
1
1
8
3.2.2. Covering the Residual Load
The residual load is the load demand that has not been covered by the PV and wind sources. We
include this method to cover the load by the generated PV or wind found in the chosen pixel that
covers the largest share of this load in comparison with other pixels.
Equation 9 describes the solar energy fraction added to reduce the residual load each time the
method is employed after each loop8. Of course, this applies to the wind fraction added when
wind is favored for installation at a certain step.
DA _ PV
normalized PV time series energyadded
9
DA_PV refers to the solar fraction added, in [MW], and the normalized PV power generation
time-series is the normalized PV power generation time-series [per unit], and the energy added
8
A description of the simulation model used is discussed in the following chapter
6
3. Technical Background
or the step size is a pre-defined value for the available PV or wind potential, which allows the
installed capacity to increase by a certain value [MW].
Where for each hour we check if the normalized PV power is bigger than the load. If this power
is bigger than the load then DA_PV would equal the load, and if not, the DA_PV equals the
normalized PV power.
This procedure applies on the different locations or regions in the country, these pixels which fall
just under the residual load curve are considered for further processing whilst these areas that fall
above the residual load curve (apparently exceeding it) would not be assigned for further
processing in this method, but would however, be considered in the third method, namely,
covering the residual load peaks.
3.2.3. Covering the Residual Load Peaks
The goal of this method is to cover the load profile peaks by installing PV and wind capacity in
locations where they best cover the load peak to limit or reduce the needed backup capacity.
We first establish the limit which defines a residual load peak, which in this investigation, is less
than 10% of the residual load demand peak, Figure 2. After that, hours are determined for which
this residual load peaks occur, and values of each residual load peak are recorded.
Finally, the best load peaks covering PV or wind normalized power time-series are used to cover
this residual load peak by a certain step size as discussed earlier.
4
x 10
7.5
7
Residual load [MW]
6.5
6
5.5
5
4.5
4
3.5
3
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
Time [Hours]
Figure 2: Residual load for Germany with 10% over Maximum Residual Load Limit
7
3. Technical Background
3.2.4. The Suitability Factor
.
Equations 10 and 11 describe the PV and wind sources‟ „suitability factor‟, which is the selection
criteria defined as the multiplication of the three methods outputs raised to a suitable power9.
At first all the three factors are normalized by the respective maximum to adjust their influence
in respect to the suitability factor, while the +1 value number added to the CC_PV 1 is added
before it is normalized, so that we get only positive values.
The suitability factor is used to weigh the contribution of the three upper mentioned methods,
and in so doing, decides whether a PV or wind capacity shall be installed, based on a higher
„suitability factor‟ value between both.
Suitability factorPV
(
CC _ PV 1 PowerCC _ PV DA _ PV PowerDA _ PV
)
(
)
MaxCC _ Pv
MaxDA
10
DA _ LS _ PV PowerDA _ LS
(
)
MaxDA _ LS
Suitability factorW
(
(
CC _ W 1 PowerCC _ W
DA _ W
)
(
) PowerDA _ W
MaxCC _ W
MaxDA _ W
DA _ LS _ W
) PowerDA _ LS _ W
MaxDA _ LS _ W
11
9
The power values were chosen arbitrarily for each term. However, they are representative of the significance of
each method to the suitability factor value, and moreover, they could be altered based on a more scientific reasoning.
8
3. Technical Background
Table 1 describes the 18 terms used in the suitability factor equations for PV and Wind.
Terms used in Equations 10 and 11
Correlation coefficient matrix for all pixels and
classes of tilt and orientation angles
CC_PV
Overall maximum correlation coefficient of
PV‟s correlation coefficient matrix
Power of the first term for PV suitability factor
Residual load cover matrix for all pixels and
classes of tilt and orientation angles
Overall maximum residual load cover
percentage of PV‟s residual load cover matrix
Power of the second term for PV suitability
factor
Peak residual load cover matrix for all pixels
and classes of tilt and orientation angles
Over all maximum cover percentage of the
residual load peaks
Power of the third term for PV suitability
factor
MaxCC_PV
PowerCC_PV
DA_PV
MaxDA
PowerDA
DA_LS_PV
MaxDA_LS
PowerDA_LS
Correlation Coefficient matrix for all pixels and
chosen hub heights
CC_W
Overall maximum correlation coefficient of
Wind‟s correlation coefficient matrix
Power of the first term for Wind suitability
factor
Residual load cover matrix for all pixels and
MaxCC_W
PowerCC_W
DA_W
chosen hub heights
MaxDA_W
PowerDA_W
DA_LS_Wind
MaxDA_LS_W
PowerDA_LS_W
Over all maximum residual load cover
percentage of Wind‟s residual load cover
matrix
Power of the second term for Wind suitability
factor
Peak residual load cover matrix for all pixels
and chosen hub heights
Over all maximum peak residual load cover
percentage of Wind‟s peak residual load cover
matrix
Power of the third term for Wind suitability
factor
Table 1: Terms Used in the Suitability Factor Calculation for PV and Wind
9
4. The Model
4. The Model
The model on which the results and analysis are based, was developed with the Matlab program.
A description of the data employed, the inputs and assumptions necessary for the model, the
equations used, and the algorithm will be presented in this chapter.
4.1. Data Used
4.1.1. Input Data
Before presenting the input data used in the model, it is important to present the concept of a
pixel, which we use when we refer to a point on the map of Germany. Each pixel therefore, is a
coordinate that defines a geographical location. In Germany, 3027 pixels are numbered and used.
Each pixel equates an area of approximately 140 km2.
The following input data is provided in matrix form to the program and a range of different
matrix sizes exists which are shown between brackets for each input:
Coordinates of each pixel in Germany (longitude, latitude) [3027 x 1].
Time-series of electrical load demand of Germany [8760 x 1].
Potential of PV sources per pixel [3027 x 1].
Wind potential per pixel [3027 x 1].
Normalized time-series for PV generation [8760 x 3027 x190].
Normalized time-series for wind generation [8670 x 3027 x 1].
The population of Germany per pixel [3027 x 1].
The input data was re-arranged into six regions of Germany, which will be later presented in the
following section of this study.
Region number
1
Number of pixels
491
2
679
3
409
4
406
5
582
6
460
Coordinates limits
Longitude (6.3125 to 10.4375)
Latitude (47.3750 to 49.8750)
Longitude (5.9375 to 10.4375)
Latitude (50.000 to 52.3750)
Longitude(6.6875 to 10.4375)
Latitude(52.5000 to 55.0000)
Longitude(10.5625 to 13.8125)
Latitude(47.3750 to 49.8750)
Longitude(10.5625 to 15.0625)
Latitude(50.0000 to 53.1250)
Longitude(10.5625 to 14.4375)
Latitude(53.1250 to 54.6250)
Table 2 Academic Division of the Regions in Germany
10
4. The Model
4.1.2. Wind, Solar and Load Demand Data
The use of the most complete and up-to-date data available is vital to the accuracy of the
simulation results. In this section, the compilation of wind, solar and load demand data used in
the model is described in some detail.
4.1.2.1.
Wind Data
The normalized (per unit) time-series generation curves were provided in one matrix of [8760 x
3027]; for 8760 hours and 3027 pixels for 2009. A sample of this normalized wind time-series
generation curve combined with the PV generation at selected tilt and orientation angles is
shown in Figure 3.
Wind Power vs. PV Power - 12 months
pu
1
1
0.5
0
0.5
0
1000
2000
3000
4000
5000
6000
7000
8000
0
9000
Time (hours)
Figure 3: PV vs. Wind Generation in Pixel One
11
4. The Model
In this research, the technical potential of wind energy in Germany is used, the data showed all
the locations (coordinate points) in the map of Germany which had a potential of 5 MW, and
these data were later refitted to suit the 3027 points used in this model.
4.1.2.1.1.
Wind Power Normalized Time Series Simulation [7]
A physical model is the basis for the simulation of power from wind turbines, and calculations
are based on stored performance characteristics and wind speeds time series.
Wind speeds are derived from the COSMO-EU model (which is a numerical weather prediction
system for the German weather service), and have the same pixel spatial resolution of [1/8° x
1/8°], with a temporal resolution of one hour.
Taken that a neutral stratification is present, wind speeds are taken at hub height and a
logarithmic wind profile is thus used. Other effects such as shading by other adjacent wind
turbines are taken into account.
4.1.2.2.
Solar Data[8]
4.1.2.2.1.
Technical PV Potential
PV system components and nominal capacity were evaluated and compiled, from different sites,
and were matched with the zip code (address) from the sites were they were collected.
PV potential on roofs and facades was identified for residential and industrial zones based on the
zip code of the site, and their calculation based on the Information coordination on the
Environment project CORINE[9] data (See Figure 4). Satellite images were used with a spatial
resolution of [100 x 100 m] where linear objects such as roads, rivers, railways, power lines and
even small villages were not detected.
Moreover, certain deviations from reality are expected, such that potential for protected areas
was calculated and included.
PV potentials along highways and railways were calculated using a model called the basic digital
landscape model (Base DLM) describing the topographical features of the landscape.
The data was finally collected by the individual German states and represent the highest spatial
resolution possible.
12
4. The Model
.
Figure 4: PV Potential According to the CORINE Land Cover Data
4.1.2.2.2.
PV Power Normalized Time-series Simulation
Global radiation data were derived from satellite images for 2009 and were used to simulate the
PV power normalized time-series, and power losses due to heating of the modules were taken
into account.
The spatial resolution of the simulated data was the same for pixels, i.e. [14 x 10] km2, with a
time resolution of one hour.
The different configurations of tilt and orientation angles were taken into account, and a standard
Polycrystalline module was used and justified because this PV module has a share of about 60%
of the installed modules.
Moreover, considering different module types was not justified since the impact on performance
was marginal.
Figure 5 shows the simulated technical potential data that was used for wind and PV in the
model (compare the technical potential of PV in Figure 5 with Figure 4).
13
4. The Model
Wind Potential
55
800
54
700
600
52
500
51
400
MW
Latitude
53
50
300
49
200
48
100
47
5
6
7
8
9
10
11
12
13
14
15
16
0
Longitude
PV Potential
800
55
Latitude
53
600
52
500
51
400
50
300
49
MW
700
54
200
48
100
47
5
6
7
8
9
10
11
12
13
14
15
16
Longitude
Figure 5: PV and Wind Potential in Germany
4.1.2.2.3.
Load Demand Data
Load demand data of Germany was provided for the year 2009, and then, values of population
per pixel were also provided and later used to scale the load demand data according to the region
for which the simulation is run. The load demand curve is shown in Figure 6.
14
4. The Model
7.5
x 10
4
7
6.5
Load demand [MW]
6
5.5
5
4.5
4
3.5
3
2.5
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
Time [Hours]
Figure 6: Load of Germany in 2009
4.2. The Program Algorithm
As can be traced from the chart flow in Figures 7 and 8, the program starts with defining several
assumptions and uses available sets of data that correspond to the specific objective sought to be
achieved; therefore, these values are fixed throughout the run time of the program.
Other scenarios could be formulated based on different assumptions and the availability of data,
rather than those presented in the following table, for example, changing the total capacity to
install or using other normalized power curves for wind at different elevations.
Table 3 describes these assumptions and a range of corresponding values used in the simulation,
the last column states if the value has been changed in the final results generation.
The values presented here are the ones used to run the simulation of Germany as a one region
model that will be discussed in the following chapter. The same assumptions were also used for
the simulation of Germany divided to regions, but certain values were changed depending on the
region under study, for example, the capacity to install or load in the region…etc
15
4. The Model
Assumption for
Total capacity to install [MW]
Power for weighing energy allocation
methods:
PowerDA
PowerDA_LS
PowerCC
Installation loop of PV and Wind
segments
Steps for installation [MW]
Year chosen for data
Hub height for wind turbines
PV orientation angles
PV tilt angles
Upper relative share of the residual
load
PV potential [MW]
Wind potential [MW]
Value
Germany: 237 908
Regions: Table 7
1
1
2
Each top 20 pixels
40
2009
100m
From -90 to +90
From 0 to +90
10%
Germany: 379.7
Regions: Table 7
Germany: 94.76
Table 3: Assumptions and Pre-Set Values Used in the Model
After these assumptions and values are set and scenario defined, the program starts with
generating the first set of different classes (combinations) of tilt and orientation angles, and
getting the 19 x 10 = 190 possible combinations.
These correspond to the normalized PV time-series for which the program calculates the three
allocation methods, which are the terms for the suitability factor formula discussed previously in
Equations 10 and 11, and were the same applies to the normalized wind time-series.
These suitability factors are calculated and compared with one another to check for the highest
value (the more suitable) each time for all installation steps in the program.
This highest value equals the maximum suitability factor of PV calculated over all pixels, and all
tilt and orientation combinations. For wind, it equals the maximum suitability factor calculated
over all pixels and for the chosen hub height of 100m.
To compare between the highest value of PV and wind, the following criteria is used:
Is MaxPV MaxWind
12
Since we have 20 steps per loop (investigating the best 20 locations suitable for the installment
of wind or PV capacity), then this condition gets tested 20 times in a loop, and then per each
16
4. The Model
answer, i.e.: MaxPV or MaxWind, a corresponding PV or wind segment gets installed, of 40
MW (or whatever step size is taken).
If MaxPV condition gets satisfied, the pixel with maximum suitability is used, and its
corresponding tilt and orientation angle recorded, and then, the generation curve of this max tilt
and orientation combination is loaded and the maximum generation curve extracted and used to
downsize the residual load:
Newresidualload previousresidualload generationcurveto use step size
13
The same applies for the MaxWind condition when it is satisfied.
At the end of this step, a test is run to see if there is further capacity on the pixel that was used, in
which if the unused capacity is less than the step size the pixel‟s value is put to zero, otherwise,
the value of the suitability factor of the used pixel with its corresponding (combination of tilt and
orientation or hub height) is set to zero meaning that it could be used again but not for the same
combination that was chosen before (in case of PV). And the loop variable capacity step, PV
segment, Wind segment gets increased correspondingly.
17
4. The Model
Real potential for each region, can be
extracted from studies or assumed at an
appropriate value
PV
Capacity in
that region
Wind
Capacity in
that region
Scenario defined (Region chosen for
analysis, year for which data are
extracted, objective setting: Running
until the mean residual load equals
zero or until a certain capacity is
installed)
Load in that
region for
that year
Angles used for
PV : Tilt (0-90) &
Orientation (-90 +90)
Wind turbine hub
heights chosen
(100 m, 50 m, 10
m...etc)
Capacity
allocation
methods and
weighing
factors
Region‟s location
(Latitude,
Longitude)
Creating a set of classes
of the different
combinations of tilt and
orientation angles
Figure 7: Main Inputs and Assumptions Used in the Model
18
4. The Model
Run equation-based functions and create
matricies for the different allocation
methods:
CC: Correlation coefficent method
CD: Covering the load method
CDP: Covering the load peaks method
PV generation
timeline per
location, tilt and
orientation angles
Maximum
CC for PV
Maximum
CCD for PV
Maximum
CDP for PV
Wind generation
timeline per
location and hub
height
Maximum
CCD for
Wind
Maximum
CC for Wind
Maximum
CDP for
Wind
Suitability
factor for
Wind
Suitability
factor for PV
Higher PV or Wind
suitability factor
PV
W
Choose best pixel of
highest suitability
factor
Choose best pixel of
highest suitability
factor
Reduce the residual load
by adding the timeline of
PV generation multiplied
by the pre-selected step
size
Reduce the residual load
by adding the timeline of
Wind generation
multiplied by the preselected step size
Save results for the
new residual load
and further
parameters
Save results for the
new residual load
and further
parameters
Figure 8: Model Steps to Simulate the Results
19
5. Analysis of Germany as One Region
5. Analysis of Germany as One Region
In order to implement the model described in the previous chapter and analyze the situation in
Germany, certain assumptions have to be carefully made and values selected. The reason for this
is the excessive run-time of the used code with a personal computer of a high speed and memory
capacity which would last for approximately 4.510 days each time the program is run.
The objective in this chapter was to draw the baseline scenario for Germany without resorting to
a location-oriented refitting, which will be the subject of the next chapter. This baseline scenario
should give us an indication of the suitable sites for the installment of the available PV and wind
capacity, and the degree to which the residual load is reduced by the installed capacity. The
scenario shows the change affecting figures such as the correlation coefficient and the suitability
factors.
In this simulation only half of the available capacity of the sum of both PV and wind was used,
allowing us to use the better-half locations for allocating PV and wind plants.
The following values and assumptions in Table 4 were used to get the finalized results for the
analysis of Germany as one region presented in this chapter.
Germany
Number of Pixels (Area covered)
Load [TWh/year]
PV Potential [GW]
Wind Potential [GW]
PV : Wind
3027 (around 348.672 km2)
459.74
379.70
94.76
≈4:1
Table 4: Percentages of Load Coverage in Germany
As demonstrated in Table 4, the PV technical potential in Germany is four-times higher than that
of wind. Matching should thus be made between where this potential is, and where it is suitable
to install the different capacity available. For example, the installed wind and PV capacity by the
end of 2010 were 27, 17.3 GW respectively (1.5:1 ratio).It would be interesting to see if the
simulation results match this ratio.
Of course, with giving a bigger weight in the suitability factor formula for a certain method over
another, this ratio could change. In this study, a higher weight was given for the allocating based
on the correlation coefficient method as the model tries to put significance on this method for its
smoothing effect consequences.
10
This value is up to ±20% correct
20
5. Analysis of Germany as One Region
5.1. Simulation Results
In the following table, some results for the simulation are shown:
Prevailing end state
Leading Source
Half
≈
37%:63%
1987
PV:Wind Mix of Annual
Generation TWh/year
100 %
Residual load ≤ 0 [Hours]
Percentage Wind installed
37.45%
65%
Percentage PV installed
94.755
Percentage of Residual
Load Reduced
Total installed wind
capacity
94.755
52,482
Capacity available Wind
[GW]
142.55
Average Load [MW]
Total installed PV
capacity [GW]
379.70
18,366
Capacity available PV
[GW]
474.460
Average Residual
Load[MW]
Total capacity available
[GW]
Table 5: Simulation Results for Germany as One Region
In this simulation, half of the total available capacity from PV and wind was installed. We see
that only 37.45% of the available PV capacity was installed, whereas the percentage is100% for
wind.
We also see that the PV share of the total generated energy of 298.850 TWh/year annually is
111.58 TWh/year or 37.33%. Then the PV:wind share of the whole generated capacity is
approximately 37%:63%.
Moreover, it is important to assess the best locations for wind allocation by studying half the
installed wind capacity, and these locations are shown later in Figure 14.
The residual load reduced to an average of 18,366 MW, a reduction of ≈ 65% from the 52,482
MW average it started with.
We note that the hours in which the residual load was below zero were 1987 hours and this can
be seen also from the Residual Load Duration Curve (RLDC) shown in the following Figure 9,
which indicates that:
The average residual load is 48,220 MW for almost 1500 hours/year (≈ 17% of the time)
The average residual load is 22,880 MW for almost 4980 hours/year (≈ 57% of the time)
The average residual load is -19,870 MW for almost 1980 hours/year (≈ 23% of the
time)
21
5. Analysis of Germany as One Region
4
8
Residual Load duration Curve for Germany
x 10
6
Avg. [MW]
= 48220
4
2
0
MW
Avg. [MW]
= 22880
-2
-4
Avg.[MW] =
-19870
-6
-8
-10
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
Time [Hours]
Figure 9: RLDC for Germany
5.1.1. PV Capacity Installed
As seen in Figure 10, much of the capacity has been concentrated in the western and southern
regions rendering them the best location for PV installment according to the installment methods.
The maximum capacity installed was ≈784 MW in pixel number 1335, which corresponds to
(7.1875 Lon. 51.5 Lat.) located in the western region.
The north-eastern region is almost empty, rendering it a highly non suitable region for PV
installment, with only a few locations where capacity doesn‟t exceed 100 MW.
The total PV capacity installed was 142,550 MW, which is 37.54% of the total available PV
capacity in Germany.
22
5. Analysis of Germany as One Region
Installed PV capacity for Germany
56
55
700
54
600
53
52
400
MW
Latitude
500
51
300
50
200
49
100
48
47
4
6
8
10
12
14
16
0
Longitude
Figure 10: PV Capacity Installed for Germany
5.1.2. PV Capacity Installed Tilt Angles and Orientation Preferences
It is important to study the ideal orientation and tilt angles, and to compare these with the actual
situation in PV plants in Germany. This will give an indication of the compatibility of the
findings currently in place in Germany, as seen in Figure 12.
In Figure 11, it appears that the tilt angle of 70° and the orientation of 60° have the highest
installed capacity. This is additionally shown in Figure 13, and compared with the actual
situation in Germany.
23
5. Analysis of Germany as One Region
Used tilt and orientation angles for Germany
x 10
100
2
90
1.8
80
1.6
1.4
60
50
1.2
40
1
30
0.8
20
0.6
10
0.4
0
0.2
80
60
40
West
20
0
South
-20
-40
-60
East
-80
MW
Tilt Angle (Inclination)
70
-10
100
4
-100
Figure 11: Tilt and Orientation Angles Used to Allocate PV Panels
0.5
Simulation results
Actual situation in Germany
Frequency
0.4
0.3
X: 22.5
Y: 0.1624
0.2
X: 70
Y: 0.4651
X: 32.5
Y: 0.2198
X: 47.5
Y: 0.1204
0.1
0
0
10
20
30
40
50
60
70
80
90
Tilt
0.5
X: 10
Y: 0.4745
Frequency
0.4
0.3
X: -10
Y: 0.09066
X: -60
Y: 0.08894
0.2
X: 60
Y: 0.2007
0.1
0
-80
-60
-40
-20
0
Orientation( W S E
20
40
60
80
)
Figure 12: Frequency of Tilt and Orientation Angles Use of the Different Installed PV Capacity
It is apparent that there is a difference between the actual situation in Germany for both the
orientation and tilt angles in respect to their ideal positioning according to the simulation.
24
5. Analysis of Germany as One Region
The results suggest an optimal tilt angle of 70°, and an ideal orientation angle of 60°, -10°, and 60°, in order of preference. The actual situation favors the tilt value of 32.5° as the maximum, as
well as 47.5° and 22.5°. As for the orientation angle, the actual situation is obviously favoring
10° angle.
5.1.3. Wind Capacity Installed
Installed Wind capacity for Germany
56
800
55
700
54
600
53
400
51
300
50
49
200
48
100
47
MW
Latitude
500
52
4
6
8
10
12
14
16
0
Longitude
Figure 13: The Installed Wind Capacity
Wind capacity is predominantly installed in the northern western and northern eastern regions,
with very low capacity installed in the south east.
The maximum capacity installed is 855 MW, at (Lon. 8, 9375 Lat. 54). However, this is not an
accurate indication of the best locations for the installment of the capacity. Because all wind
capacity has been installed we have to check for the best locations where half of the wind
capacity (47,377 MW) was installed and see if these are the same locations that are proposed in
Figure 13.
We can also decide on another value specified from Figure 15, like the step value at which the
slope of installed wind capacity begins decreasing.
25
5. Analysis of Germany as One Region
The new figure for half the capacity installed is accordingly shown in Figure14, white regions
within the map represent those pixels with no wind potential.
Percentages of the installed capacity to the available capacity are used to give an indication of
the installation order at half the installation steps of the full capacity.
Finally, Differences between installing the 47 GW to the available 94 GW are clear. For
example, the maximum new capacity installed in this second case is 520 MW at the pixel
number 282 (Lon. 7,4375 Lat. 53,625).
Half installed wind capacity for Germany
56
100
55
90
80
54
70
53
50
[%]
Latitude
60
52
51
40
50
30
49
20
48
47
10
4
6
8
10
12
14
16
0
Longitude
Figure 14: Half the Installed Wind Capacity for Germany
5.2. Discussion and Conclusion of the results
It is obvious from comparing Figure 5 to Figure 14 that all the available capacity from wind has
been installed, highlighting the importance of differentiating at what point the differentiation
between PV and wind stopped, as shown in Figure 15.
In Figure 16, the load demand is shown and plotted against the generation from both PV and
wind capacity. The first plot in the upper location has the whole year time scale, i.e. 8760 hours.
It is apparent that the load is not well covered in the first 1500 and last 1000 hours, the winter
months in Germany, and that it is satisfactorily covered in other times, such as between 20003000 hours. The first observation is shown in the middle plot for the first 1000 hours (until
around Mid-February). We can see few hours where peaks and loads are covered. By looking
26
5. Analysis of Germany as One Region
only at the plot, we can also see that good correlation is not prevalent although load peaks and
generation peaks are more or less correlated together.
15
x 10
4
PV vs. Wind installation order for Germany
PV
Wind
MW installed
10
Wind
capacity
extinguished,
PV
available
only
5
Wind
capacity
available
0
0
1000
2000
3000
4000
5000
Intallation Steps
6000
7000
8000
Figure 15: Installation of PV and Wind Capacity
In the last plot of Figure 16, which is between 2200-3000 hours, better results for load and load
peaks coverage as well as the overall correlation coefficient are evident. We would expect these
hours to have the better share of the below-zero residual load total at the end.
In Figure 17 we can see the overall change of the three allocation methods with installation steps.
Several factors influence the shape of the decreasing suitability factor plot. For example, the
sudden increase of load peaks coverage and better correlation with the load along rather constant
load coverage after around 200 steps, high suitability factor prevails.
Keep in mind that these maximum values occur only for the first step in each loop which is
defined by assigning the best top 20 locations, before next round of allocations, so sudden drops
or changes in these trends should be expected.
Finally, we notice that the maximum residual load has not significantly decreased, only about
≈ 7% of the maximum residual load was decreased, but we notice that the average residual load
has significantly decreased by about 65%. There could be different reasons for that;
This might be caused by the low weighing of the coverage the loads peak method.
Might be due to certain particularly high peaks that could not be covered because of the
non-existent PV-wind generation at the respective time-frame, so that the peak would
never reduce by any of the allocation methods.
Next, we examine the regional results and identify the differences that exist in the overall
residual load coverage and capacity installation when regions are created from the division of
Germany.
27
5. Analysis of Germany as One Region
15
x 10
4
Load demand vs.capacity generation from PV-Wind share
MW
10
5
0
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
Time[Hours]
x 10
4
15
Capacity generation
Low correlation
Load demand for Germany 2009
MW
10
5
0
0
100
200
300
400
500
600
700
800
900
1000
Time[Hours]
x 10
4
10
MW
Peaks covered Load coverage
5
0
2300
2400
2500
2600
2700
2800
2900
Time[Hours]
Figure 16: Different Time Ranges for Residual Load versus Generation Capacity for Germany
28
5. Analysis of Germany as One Region
Correlation Coefficent - Germany
-1 < CC < 1
0.5
Wind capacity extinguished
0
-0.5
-1
0
1000
2000
3000
4
MW
15
4000
5000
6000
7000
8000
Load Coverage - Germany
x 10
10
Wind capacity extinguished
5
0
0
1000
2000
3000
4000
5000
6000
7000
8000
5000
6000
7000
8000
6000
7000
8000
6000
7000
8000
Load Peaks Coverage - Germany
MW
4000
2000
0
0
1000
2000
3000
4000
0 < suitability < 1
Suitability Factor - Germany
1
Wind capacity extinguished
0.5
0
0
1000
2000
4
MW
7.5
3000
4000
5000
Maximum Residual Load - Germany
x 10
7
6.5
0
1000
2000
3000
4000
Installation steps
5000
Figure 17: Allocation Methods Change for Germany and Suitability Factor Decrease in Analyzing These Different Results
29
6. Analysis of Germany Divided into Regions
6. Analysis of Germany Divided into Regions
New qualitative predictions can be deduced when dividing the whole set of Germany‟s input data
to newly created regions within it.
Six regions were hence developed to answer this study objective. Six regions present a random
selection that is suitable for many considerations. It shows the difference between the western
and eastern regions which have population density differences, Figure 18. The variation of PV
and wind sources availability change also from north to south, Figure 5. And such a division has
an ability to naturally accommodate an argument for further variances unaccounted for.
These regions mainly represent six known geographical directions: south-west or region one,
mid-west or region two, north-west or region three, south-east or region four, mid-east or region
five, and north-east or region six.
The objective thus for this division is to investigate the placement of capacity shift between the
established regions in comparison with the distribution of these capacity for Germany as one
region. This uncovers qualitative hints regarding storage capacity change and the needed
transmission capacity.
After that, we would investigate how to limit the storage and transmission capacity. It is useful to
have these regions because that allows us to study if the excess energy in any region can be
reduced to compensate the losses that would occur due to the selection of the poorer sites, as in
comparison with the Germany as one region scenario and with a reference scenario, that presents
the expansions expected under the current conditions. Furthermore, the results can give us
answers for such questions that may also arise:
Which locations exhibit the highest order for capacity installment...from wind sources…from PV
sources? And how do the three allocation methods compare with each other in each of the
regions...and with other regions?
And one may also ask: which regions are able to be a future energy exporter of energy to bring
the residual load of Germany to a minimum...and which will act as an energy importer?
These and many other questions are certainly interesting to answer, such as the establishment of
a national Direct-Current (DC) grid.
To answer questions concerning transmission and storage a basic transmission network model
for energy exchange between the regions and a model of storage for the energy time shift is used
section 6.2.
30
6. Analysis of Germany Divided into Regions
Table 6 describes numerically the number of pixels (or locations) in each region, the respective
area, the population size, the PV and wind potential, and the load demand in each region. This
load demand is linearly dependent on the population size.
Three indicators for PV and wind output are shown for half the used potential in each region:
half capacity PV output, and half capacity wind output. Finally, the ratio between the full
potential of PV to wind is presented.
Number of
Pixels
Area [Km2]
Population
PV
Potential
[GW]
Wind
Potential
[GW]
Load
[TWh/year]
PV : Wind
(potential)
Reg. 1
Reg. 2
Reg. 3
Reg. 4
Reg. 5
Reg. 6
491
679
409
406
582
460
56,557
17,503,702
78,212
28,365,287
47,117
8,644,355
46,766
9,490,845
67,039
11,308,329
52,986
7,050,544
63.2
112.0
47.7
40.6
77.7
38.7
4.785
21.235
28.780
0.835
20.170
18.155
97.7
158.3
48.3
53
63.1
39.4
13.2:1
5.27:1
1.66:1
76:1
3.85:1
2.13:1
Table 6: Percentages of Load Coverage in Germany by the Six Newly Established Regions
In studying Table 6, a first rough estimation of the situation in the newly established regions is
predicted. There is a high PV to wind potential that varies between the regions. However, this
alone cannot explain to us the full picture; what we need to know is how much of this available
potential from both PV and wind is optimally installed based on the energy allocation methods.
Also, we need to know which of the two sources will participate more in the 100% renewable
share, and finally, whether this capacity will be able to cover the load demand at that region.
In the last row, we can see the ratio of the available PV potential in each region to the available
wind potential. What we care for in this, is how much of this potential is actually suitable to
optimally cover the load demand curve.
In Figure 18, the regions with the highest densities are predominantly in the south-west (region
one) and mid-west (region two). This will have an effect on the coverage of the residual load in
these regions if the sources available are not enough.
31
6. Analysis of Germany Divided into Regions
54
53.5
5
54.5
10
54
53.5
53
5
53
7
8
9
10
52.5
10
11
11
Longitude
5
50.5
Latitude
51
Inhabitants
Latitude
4
x 10
15
52
10
51.5
10
51.5
51
5
50.5
5
6
7
8
9
10
50
10
11
11
12
Longitude
10
48
5
8
9
Longitude
10
11
Latitude
49
Inhabitants
50
7
14
15
16
South-East (Region 4)
4
x 10
15
6
13
Longitude
South-West (Region 1)
Latitude
15
52.5
52
11
14
Mid-East (Region 5)
4
x 10
15
52.5
47
13
Longitude
Mid-West (Region 2)
50
12
Inhabitants
6
4
50
x 10
15
49
10
48
5
47
10
11
12
13
Inhabitants
52.5
4
x 10
15
55
Latitude
10
Inhabitants
54.5
Latitude
North-East (Region 6)
4
x 10
15
Inhabitants
North-West (Region 3)
55
14
Longitude
Figure 18 the Six Regions with Population Densities in Colored Code11
The maps were rescaled and show population densities up to 150,000 inhabitants per location
32
6. Analysis of Germany Divided into Regions
6.1. Simulation model
6.1.1. Experimental framework
The six regions were created by allocating each set of data to its corresponding region, then, the
main program designed for the analysis of Germany as one region was refitted to calculate the
same results for each region.
Several new additions to the main algorithm were made to both optimize the results for Germany
as one region and the six regions developed: The real potential of PV and wind plants from two
different studies12 was used. Different trials were executed with different assumptions throughout
the course of the research. These included using different step sizes (10MW, 25MW, 40MW,
50MW), the criteria was to have a realistic installation capacity, that‟s able to distinguish the
results for the optimal locations instead of using all the available locations, hence, a 40 MW
installation step size was selected as the best choice.
This might have reduced the quality of the installed capacity allocation at a time but was
necessary nonetheless to speed up the processing of the data by approximately 75%. This simply
means that we install a plant of 40 MW at a suitable location instead of 10 MW or 50 MW,
which is an acceptable size if our objective is a gross installation of wind and PV farms with high
generation capacity.
The assumptions made for each region, are the same used for the treatment of Germany as a one
region. The same data were also used, but by allocating each set of data to its corresponding
region as mentioned earlier.
Due to the large size of data dealt with, different combinations, new sets of data that needed to be
generated, and functions solved, processing time for each region took between 8-12 hours at a
time and hence, minimizing processing time and having a proper memory size was a concern for
the successful execution of the program with Matlab.
Finally, after the program is run and results simulated, two states at which the program stops
running are: either half the available capacity is installed and the average residual load is less or
equal to zero, or that half the available capacity is installed but the mean residual load is still
bigger than zero.
12
These studies are cited in the references number 7 and 8
33
6. Analysis of Germany Divided into Regions
6.1.2. Simulation results
In Table 7, a summary of the main results is shown.
Total capacity
available [GW]
Total installed
capacity [GW]
Capacity
available PV
[GW]
Total installed
PV capacity
[GW]
Capacity
available Wind
[GW]
Total installed
wind capacity
Percentage PV
installed
Percentage Wind
installed
Average Residual
Load[MW]
Average Load
[MW]
Percentage of
Residual Load
Reduced
Residual load ≤ 0
[Hours]
PV:Wind Mix of
Annual Generation
TWh/Year [%]
Leading Source
End state of
installation
South-West
Reg. 1
67.98
Mid-West
Reg. 2
133.22
North-West
Reg. 3
76.44
South-East
Reg. 4
41.37
Mid-East
Reg. 5
97.83
North-East
Reg. 6
56.86
34.01
66.61
38.76
21.13
49.07
28.43
63.16
111.98
47.66
40.54
77.66
38.71
29.19
45.88
11.96
20.29
28.90
11.37
4.76
21.24
28.78
0.84
20.17
18.16
4.76
21.24
28.78
0.84
20.17
17.56
42.93%
40.97%
25.10%
50.05%
37.21%
29.36%
100%
100%
100%
100%
100%
96.69%
7699.39
8256.32
-2174.58
3830.37
-456.26
-970.62
11153.00
18074.00
5508.00
6047.60
7205.70
4492.60
30.96%
54.32%
139.50%
36.66%
106%
121.6%
643
1417
4894
1249
3819
4504
77%:23%
47%:53%
16%:84%
91%:9%
38%:62%
22%:78%
PV
Half the
total
capacity
installed
PV
Half the
total
capacity
installed
WIND
Average
residual
load less
than zero
PV
Half the
total
capacity
installed
WIND
Average
residual
load less
than zero
WIND
Average
residual
load less
than zero
Table 7 Average Residual Loads and the Leading Source in Each Region after Installing Half the Available
Capacity
The capacity available is the sum of all the available PV and wind provided in Table 6, where
approximately, half of the available capacity was used as a criterion to distinguish properly the
ideal top 50% of the available locations selected for capacity installment.
34
6. Analysis of Germany Divided into Regions
Following upon the experimental framework discussion, the end state of installation was average
residual load less than zero (first end state) for regions 3, 5 and 6 and half the total capacity
installed (second end-state) regions 1, 2 and 4.
We can also notice that it is mainly in these regions of high wind capacity installment where it is
a leading source that we have the first end-state satisfied. The effect of installing PV capacity in
regions where it is the leading source is mainly because wind capacity is not enough in that
region and that‟s when the second end-state prevails.
The percentage of residual load reduced is a measure of how well did the installed capacity
manage in reducing the load demand in the respective region. The total new residual load is
141.773 TWh/Year for all regions, thus, the overall reduction in the residual load is 69.1%
The leading source, gives us a sense of the predominant capacity type to be installed for the
attainment of the stated reduction in the load demand in the respective region. However, it is
seen clearly that wind prevails as a total sum of annual energy production in four of the six
regions 2, 3, 5 and 6.
It is important to note in the last line, that the two end states describe how the processing of
installing more capacity was done. First we see if the region can reach an average residual load
of zero without using half of the capacity, and if so, we go on with installing the rest of the
capacity. Otherwise, we install half of the capacity without reducing the average residual load
below zero.
Next, based on the leading source in each region, regional results for the south-west, mid.-west,
and south-east regions are presented for PV installed capacity. Also, optimal tilt and orientation
angles setting for each region are presented and a comparison with the real tilt and orientation
angles in place in Germany today made.
Finally, regional results are attained for the north-west, mid.-east13 and north-east regions for the
wind capacity installed.
6.1.2.1.
South-West region (Region One)
Following are three figures for the characteristics of the installed PV capacity in the south-west
region (region one).
13
An exception was made to add this region to the wind analysis part due to its relatively close 1:1 ratio with the PV
capacity installed and so, it would hold true for the argument in the wind section.
35
6. Analysis of Germany Divided into Regions
Installed PV capacity, South-West
50
500
450
49.5
400
350
49
250
48.5
MW
Latitude
300
200
150
48
100
47.5
50
6
6.5
7
7.5
8
8.5
Longitude
9
9.5
10
10.5
11
Figure 19: Installed PV Capacity in the South-West Region
In Figure 19 we see the distribution of the installed PV capacity, and see where the maximum
capacity was installed. For example we have a maximum of ≈ 500 MW at pixel 241 at the
coordinate point (Lon. 9.1875 Lat. 48,875). However, the capacity installed is ≈ 60 MW on
average per location.
Tilt and orientation angles, South-West Region
100
3000
90
80
2500
60
2000
50
1500
40
Times Used
Tilt Angle (Inclination)
70
30
1000
20
10
500
0
-10
100
80
60
40
West
20
0
South
-20
-40
-60
-80
-100
0
East
Figure 20: Orientation and Tilt angles for the Installed Capacity of Region One (South-West)
36
6. Analysis of Germany Divided into Regions
0.35
Simulation Results
Real Situation in Germany
0.3
Frequency
0.25
X: 22.5
Y: 0.1624
0.2
X: 70
Y: 0.33
X: 32.5
Y: 0.2198X: 40
Y: 0.18
X: 47.5
Y: 0.1204
0.15
0.1
0.05
0
0
10
20
30
40
50
60
70
80
90
Tilt
0.5
X: 10
Y: 0.4745
Frequency
0.4
0.3
X: -70
Y: 0.13
0.2
X: -10
Y: 0.11
X: 70
Y: 0.19
0.1
0
-80
-60
-40
-20
0
Orientation( W S E
20
40
60
80
)
Figure 21: Comparison between Frequency of Tilt and Orientation Angles used for the Different Installed PV
capacity in Region one (South-West)
In Figure 20 and the explanatory Figure 21, we see that the tilt angle 70° and orient angles 70°,10° and -70° were the mostly repeated angles for installation. This means that these angles are
highly suitable for the installment of the PV capacity. However, it is clear that this doesn‟t go inline with what is in place in Germany today, with most frequently 32.5° used tilt angle and 10°
orient angle.
6.1.2.2.
Mid-West Region (Region two)
In Figure 22 we see again the distribution of the installed PV capacity, and get results for the
maximum location, which is ≈ 784 MW at pixel 430, coordinate (Lon. 7, 1875 Lat. 51, 5). The
installed capacity is ≈ 67 MW per each location on average.
In Figure 23 and the explanatory Figure 24, we see that in the mid-west region, tilt angle 50° and
orient angle -10° were the mostly repeated angles for installation, meaning their high suitability
for the installment of the PV capacity. However, it is clear that this doesn‟t go in-line with what
is being used in Germany today for tilt 32.5°, but which is close enough to Orient 10°.
37
6. Analysis of Germany Divided into Regions
Installed PV capacity, Mid-West
52.5
700
52
600
51
300
50.5
200
100
50
5.5
6
6.5
7
7.5
8
8.5
Longitude
9
9.5
10
0
10.5
Figure 22: Installed PV Capacity in the Mid-West Region
Used tilt and orientation angles for Region Two (Mid-West)
100
8000
90
7000
80
6000
70
60
5000
50
4000
40
Times Used
Tilt Angle (Inclination)
Latitude
400
MW
500
51.5
3000
30
20
2000
10
1000
0
-10
100
80
60
40
20
0
-20
-40
-60
-80
-100
0
West
South
East
Figure 23: Orientation and Tilt angles for the Installed Capacity of Region Two (Mid-West)
38
6. Analysis of Germany Divided into Regions
0.5
Simulation Results
Real Situation in Germany
Frequency
0.4
0.3
X: 22.5
Y: 0.1624
0.2
X: 50
Y: 0.4309
X: 32.5
Y: 0.2198
X: 47.5
Y: 0.1204
0.1
0
0
10
20
30
40
50
60
70
80
90
Tilt
0.5
X: 10
Y: 0.4745
Frequency
0.4
0.3
0.2
X: -10
Y: 0.2303
0.1
0
-80
-60
-40
-20
0
Orientation(
20
WSE
40
60
80
)
Figure 24: Comparison between Frequency of Tilt and Orientation Angles used for the Different Installed PV
Capacity in Region One (Mid-West)
6.1.2.3.
South-East Region (Region four)
Installed PV capacity , South-East
50
700
49.5
600
500
400
48.5
MW
Latitude
49
300
48
200
100
47.5
10.5
11
11.5
12
12.5
Longitude
13
13.5
14
Figure 25: Installed PV Capacity in the South-East Region
39
6. Analysis of Germany Divided into Regions
Tilt and orientation angles, South-East Region
100
90
2500
80
2000
60
50
1500
40
30
Times Used
Tilt Angle (Inclination)
70
1000
20
10
500
0
-10
100
80
60
West
40
20
0
South
-20
-40
-60
East
-80
0
-100
Figure 26: Orientation and Tilt Angles for the Installed Capacity of Region Four
(South-East)
0.45
Simulation Results
Real Data for Germany
0.4
X: 70
Y: 0.4317
0.35
Frequency
0.3
0.25
X: 22.5
Y: 0.1624
0.2
X: 50
Y: 0.2417
X: 32.5
Y: 0.2198
X: 47.5
Y: 0.1204
0.15
0.1
0.05
0
0
10
20
30
40
50
60
70
80
90
Tilt
0.5
X: 10
Y: 0.4745
Frequency
0.4
0.3
0.2
X: -60
Y: 0.1567
0.1
0
X: -10
Y: 0.2233
-80
-60
-40
-20
X: 70
Y: 0.07167
0
Orientation( W S E
20
40
60
80
)
Figure 27: Comparison between Frequency of Tilt and Orientation Angles used for the Different Installed PV
Capacity in Region Four (South-East)
40
6. Analysis of Germany Divided into Regions
In Figure 25, we see once more the distribution of the installed PV capacity, and get results for
the maximum location, which is ≈ 778MW at pixel 97 at coordinate (Lon. 11,5625 Lat. 48,125)
,the mean capacity installed is ≈ 50 MW.
In Figure 26 and the explanatory Figure 30, we see that in the south-east region, two tilt angles
50° and 70°, and the orient angle -10° were the mostly repeated angles for installation, meaning
their high suitability for the installment of the PV capacity. However, it is clear that this doesn‟t
go in-line with what is being used in Germany today for 32.5° tilt angle, but which is close
enough to the Orient angle of 10°.
Next, we present the wind capacity for the other three regions, and naturally, we can compare
between the maximum capacity installed, and more importantly, the percentage installed shown
in Table 7.
6.1.2.4.
North-West region (Region three)
Installed Wind capacity, North-West
500
55
450
400
54.5
54
300
250
53.5
MW
Latitude
350
200
150
53
100
50
52.5
6.5
7
7.5
8
8.5
9
Longitude
9.5
10
10.5
0
Figure 28: Installed Wind Capacity in the North-West Region
In this region, the maximum capacity that was installed was ≈520 MW, at the pixel number 257,
at the coordinate point (Lon. 7, 4375 Lat. 53,625), with an average installed wind capacity of
≈65 MW per location.
41
6. Analysis of Germany Divided into Regions
6.1.2.5.
Mid-East region (Region Five)
Installed Wind capacity, Mid-East
52.5
450
400
52
350
300
250
MW
Latitude
51.5
51
200
150
50.5
100
50
50
10.5
11
11.5
12
12.5
13
13.5
Longitude
14
14.5
15
15.5
0
Figure 29: Installed Wind Capacity in the Mid-East Region
In this region, the maximum capacity that was installed was ≈455 MW, at the pixel number 439,
at the coordinate point (Lon. 13, 5625 Lat. 51,875), and the average installed wind capacity per
location is ≈ 35 MW.
6.1.2.6.
North-East region (Region Six)
Installed Wind capacity, North-East
350
54.5
300
54
200
53.5
MW
Latitude
250
150
53
100
50
52.5
10.5
11
11.5
12
12.5
13
Longitude
13.5
14
14.5
0
Figure 30: Installed Wind Capacity in the North-East Region
42
6. Analysis of Germany Divided into Regions
In this region, the maximum capacity that was installed was ≈390 MW, at the pixel number 78,
at the coordinate point (Lon. 11, 9375 Lat. 52, 75) with an average capacity of ≈38 MW per
location.
6.1.2.7.
Selected suitability factors change with installation steps
In Figure 31, results are shown for the terms of the suitability factor in the first three plots; the
correlation coefficient, load coverage, and the load peaks coverage methods. The fourth plot is
the suitability factor itself, and the fifth plot shows the maximum residual load decrease with
each installation step.
It can be seen that for the correlation coefficient (the same analysis applies for the load coverage
and load peaks coverage methods), that the values the installation began with were well above 0
0.147 and then reduced in a fluctuating trend to -0.6538.
It can also be seen that the value of the correlation coefficient seems to locally increase at some
points, which is probable; since the correlation coefficient values are renewed repeatedly per
loop or top 20 locations found before the resetting of all the variables (maximum correlation
coefficient, coverage of peaks, and coverage of residual load peaks).
In the final plot we see the sharp decrease of the residual load peaks at the beginning because of
high wind influence for peaks reduction. Then the slower rate with installation steps after that
because of the incapacity of PV to cover load peaks.
43
6. Analysis of Germany Divided into Regions
Correlation Coefficent - South West
-1 > CC > +1
0.5
0
-0.5
-1
0
200
400
MW
10
600
800
1000
1200
800
1000
1200
800
1000
1200
Load Coverage - South West
4
x 10
5
0
0
200
400
600
Load Peaks Coverage - South West
MW
6000
4000
2000
0
0
200
400
600
1 > Suitability > 0
Suitability factor - South West
1
Decreasing trend
0.5
0
0
200
400
MW
1.6
600
800
1000
1200
800
1000
1200
Maximum Residual Load - South West
4
x 10
1.55
1.5
0
200
400
600
Installation Steps
Figure 31: Allocation Methods Change for Region One and Suitability Factor Decrease in Analyzing the Different Results
44
6. Analysis of Germany Divided into Regions
6.1.3. Discussions and Conclusion of the Results of the Simulation Model
The effect of the available capacity of PV and wind on the maximum residual load reduction is
evident as well as load peaks coverage rate. Figure 32 demonstrates this. At the installation step
373 almost all the wind capacity was installed and that is when the PV installation rate became
increasingly linear and the wind capacity became constant. This situation however is not so for
regions where there is almost an equal share of PV and Wind capacity, notably, the north-west,
mid-east and north-east regions, this is shown in Figure 33.
PV vs. Wind installation order for South-West
4
3
x 10
PV
Wind
2.5
Wind capacity
available
Wind capacity
extinguished,
PV available
only
MW installed
2
1.5
1
X: 340
Y: 4756
0.5
0
0
200
400
600
800
1000
1200
Intallation Steps
Figure 32: Installation Order for PV and Wind Shares in the South-West Region (Region One)
PV vs. Wind installation order for North-East
18000
PV
Wind
16000
14000
MW installed
12000
10000
8000
6000
Wind capacity
and PV
available to
the last
installation
step
4000
2000
0
0
100
200
300
400
500
600
Installation steps
700
800
900
1000
Figure 33: Installation Order for PV and Wind Shares in the North-East Region (Region Six)
45
6. Analysis of Germany Divided into Regions
For the installed PV capacity, the tilt angles used are mainly, 70° and 50°, one reason is that with
those angles we have a higher generation in winter and a lower generation in summer in
comparison to tilt angles around 30°.
This means that the generation is more constant over the year, has better correlation with the load
and produces less excess energy in summer where a higher percentage of load is covered. This
however is quite different from the 32.5° tilt in place today in Germany.
This argument also holds true for the daily cycle where it is important not to produce excess
energy at the day peak but to try and distribute the generation all over the day period and to
match as much as possible with the load demand.
Using these optimized angles will reduce the storage needed daily and seasonally, which is an
important consideration to take into account when studying the whole supply system.
The same argument holds true for the orientation angle, where the most frequent orient angle was
-10°, which is relatively closer to the value of 10° degrees in place in Germany.
The main reasoning for this difference from the simulation results is that the simulation is based
on a suitability factor that takes into account the whole system and, primarily, the correlation
coefficient. At the other hand, it could be that planners are more interested in the bulk energy
produced to benefit from the revenue of selling this energy without any consideration to the
effect on the optimal reduction of the residual load.
6.2. The Direct Current Load Flow Matrix – Transport Capacity
To enable energy exchange between the six regions, a model of a Direct Current (DC) network is
created, that connects the regions for the transportation of power in each hour to attain an hourly
zero residual load.
It is important to consider a suitably sized storage and generator of energy at these hours in
which it is impossible to overcome the residual load.
In this regard several possibilities or assumptions are made:
1- There is a large energy storage/generator that accommodates all the excess energy in a
certain hour which can be used to generate this energy back when there is a lack of
energy in another hour.
2- There is a certain transportation capacity which is based on the optimization of the
maximum flow of energy in a certain hour.
3- A Direct Current Flow Load Matrix (DCFLM)14 determines the best path for the flow of
energy based on the physical conditions that exist in the power flow lines.
14
See Appendix A
46
6. Analysis of Germany Divided into Regions
This division is random and rather academic in nature. This means that no consideration was
given to the real network and generation capacity that is in place today. This is mainly because
the objective of this study is not to optimize the currently established grid, but rather, to provide
a sense of the optimal situation that future projects could follow and additionally provide a sense
of the gross flow requirements.
In Figure 34, we see the six regions of Germany and the configuration of the power flow lines
between the centers of the regions. Eleven lines that connect the regions together were suggested
taking into consideration that a flow must happen at an economical and technically feasible
basis. This is possible for a direct line connecting the most two adjacent centers of regions
together without passing through a center of another region.
The possibilities then for export or import (based on each hour) from Region 1 to any location
must be first to Regions 2, 5 or 4 respectively. If Region 1 is to export to region 6, this is to
happen by flowing before in the lines of the other regions. This is because it doesn‟t make sense
to create a line specifically between region 1 and 6. This argument holds true for the other
regions as well.
It must be noted however, that the DCLFM always passes a certain amount of power flow from
any region to all regions, but these quantities differ and are a characteristic of the DCLFM as will
be shown in the results later.
Regional DC Power Flow
55
14000
54
3
6
12000
53
Latitude
5
51
2
8000
50
MW
10000
52
6000
49
4000
1
48
4
2000
47
4
6
8
10
12
14
16
Longitude
Figure 34: The DC Network between the Proposed Six Regions Shown According to their Geographical
Orientation
47
6. Analysis of Germany Divided into Regions
In order to create this DCLFM, a storage location needs to be specified. Two options were
possible: the first, in the middle of the map. The problem with this option is that it is feared that
the power flow using the DCLFM will prefer a certain flow path through the storage node, which
prohibits us from taking advantage of the full exchange of power between the regions before
energy exchange between all regions.
The second option assumes a virtual storage based in a very distant location that will have the
least favorable flow of power before the balancing takes place. The second option is therefore
used. Table 8 shows the distances between the centers of the regions. The last two rows compare
between the distance to the center of the map and the other - rather fictional - large distance
storage point in the Atlantic Ocean! On which, the further analysis was based.
Regions
1
2
3
1
2
3
4
5
6
Storage (Central)
Storage (Large
distance)
0
252,5561
500,1158
221,5332
379,2037
560,1401
275,6305
2751,098
0
253,6068
374,0734
288,2075
375,7349
148,1956
2641,371
0
564,695
325,1446
231,9968
270,0271
2627,498
4
5
0
308,8203
0
529,6207
222,24
294,7206 141,8472
2966,516 2910,012
6
0
284,6167
2854,933
Storage
(S)
0
0
Table 8: Centrally Located Storage Scenario and DC Lines Distances (km)
After the optimization and shares of each source have been reached in the first step, the residual
loads of the different regions are compared with one another and the hourly and yearly storage
required calculated to have a total residual load equal to zero.
In Figure 37, the residual loads reached for each region are shown for the whole year where the
line at the 0 MW residual load indicates the overall tendency of the region to be an energy
exporter or importer. This of course can be validated later on by seeing the load flow direction
and amount between the regions.
The hourly storage equals the summation of the total residual loads of the different regions:
6
residualloadsof the six regions residualload of storage 0
11
reg 1
Figure 35 shows the residual load duration curve15 for the hourly storage capacity. By examining
the dotted red line‟s intersection with the x-axis, we see that for 2255 hours (≈25% of the year)
the regions are exporting power and for the other 75% of the time, power is being imported.
15
In section 6.2.1.1.1 a detailed definition and further examples of the residual load duration curve are given.
48
6. Analysis of Germany Divided into Regions
If we study the individual regions, we can understand where this percentage is coming from for
each region. This is shown in Figure 36 where one can see the hour percentages which are
positive.
Clearly, the higher this percentage is, the less is the region covering the residual load and the
more it will tend to import energy from other regions first before taking it from the storage.
Residual Load Duration Curve for the Regional Storage
4
12
x 10
10
6
X: 588
Y: 3.53e+004
4
2
0
-2
X: 7014
Y: -4.06e+004
-4
X: 8760
Y: -7.078e+004
-6
-8
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
Time [Hours]
Figure 35: Residual Load Duration Curve for the Regional Storage
Positive Residual Load per Hour per Region
93%
8000
86%
84%
7000
75%
6000
Time [Hours]
Residual Load [MW]
8
56%
5000
49%
50%
44%
4000
3000
2000
1000
0
1
2
3
4
5
6
Regions
Figure 36: Hour Percentage where there is a Positive Residual Load per Region
49
6. Analysis of Germany Divided into Regions
North-West (Region 3)
4
1
x 10
North-East (Region 6)
4
1
x 10
0
MW
MW
0
-1
-1
-2
-3
0
1000
2000
4000
5000
6000
7000
8000
-2
9000
Time [Hours]
Mid-West (Region 2)
4
4
3000
x 10
MW
MW
2000
3000
4000
5000
6000
7000
8000
9000
7000
8000
9000
7000
8000
9000
Time [Hours]
Mid-East (Region 5)
x 10
0
0
-2
-1
-2
0
1000
2000
3000
4000
5000
6000
7000
8000
-3
9000
Time [Hours]
South-West (Region 1)
4
2
1000
4
1
2
-4
0
x 10
0
1000
2000
4000
5000
6000
Time [Hours]
South-East (Region 4)
4
1
3000
x 10
0.5
MW
MW
1
0
0
-0.5
-1
0
1000
2000
3000
4000
5000
Time [Hours]
6000
7000
8000
9000
-1
0
1000
2000
3000
4000
5000
6000
Time [Hours]
Figure 37: Hourly Residual Loads for One Year, with Positive Residuals Above the Red Line and Negative Residuals Below it
50
6. Analysis of Germany Divided into Regions
6.2.1. The DCLFM and Load Flow Parameters
In Table 9 The DCLFM16 is defined showing the coefficients of flow in a set of buses (nodes)
and lines, in our model, 7 buses and 17 lines are defined, generating a (17 x 7) matrix, two main
inputs were provided to generate this matrix with the Matlab program, Bus and Line, moreover,
other assumptions were also necessary for that [10]:
Bus:
Number (from 0 to 6); where 0 is the storage point.
Longitude and Latitude values for where each bus is positioned.
Lines:
Number (from 0 to 6); where 0 is the storage point.
Line direction: A definition of the line from its starting bus to its target bus.
Length of lines in km
Lines
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
FromTo
1-2
1-5
1-4
1-S17
2-3
2-6
2-5
2-4
2-S
3-6
3-5
3-S
6-5
6-S
5-4
5-S
4-S
Slack
Node
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
SouthWest (1)
0,361
0,273
0,166
0,200
0,100
0,071
0,013
0,004
0,174
0,002
-0,067
0,165
-0,073
0,146
-0,010
0,156
0,160
MidWest (2)
-0,182
0,007
0,009
0,167
0,185
0,137
0,168
0,136
0,191
0,007
0,004
0,174
-0,009
0,153
0,013
0,157
0,159
NorthWest (3)
-0,097
-0,060
-0,001
0,157
-0,349
0,003
0,013
0,063
0,173
0,159
0,284
0,208
0,008
0,154
0,090
0,156
0,152
SouthEast (4)
0,014
-0,016
-0,164
0,166
0,077
0,052
-0,032
-0,255
0,172
0,000
-0,088
0,165
-0,093
0,146
-0,389
0,159
0,192
MidEast (5)
-0,006
-0,146
-0,014
0,165
0,010
0,000
-0,171
-0,017
0,173
-0,005
-0,159
0,173
-0,159
0,154
0,193
0,174
0,162
NorthEast (6)
-0,084
-0,07
-0,003
0,157
-0,011
-0,287
-0,01
0,053
0,172
-0,185
0
0,174
0,339
0,189
0,102
0,157
0,152
Table 9: DCLFM Matrix for the Six Regions for a Distant Storage Configuration
In this matrix, rows represent the lines through which the power is flowing, and the columns
represent the different regions. Several ways of reading this matrix are possible to fully
understand how it works.
16
17
More details on the DCLFM can be found in appendix A
S = Storage node
51
6. Analysis of Germany Divided into Regions
Firstly, if we look at the column South-East (region 4) for example, and sum vertically all
the absolute numbers that belong to the lines of region 4 (circled), the summation of the
absolute values should equal one (0,164+0,255+0,389+0,192 = 1). Other values in the
same column provide other possibilities for the flow from or to region 4 but for
significantly less values.
Secondly, if we take the sum of each row, which represent each line, the value is
approximately equal to zero, and if not, is set to zero by the summation with the slack
node. The exception however is that the sum of each row in which storage is an element
is equal to approximately one.
Thirdly, the positive sign in each column indicate that the flow is in direction of the line
definition as explained previously in the section.
To finally get the residual load flow between regions for each hour, the following equation is
used:
Load Flow( LF )
DCLFM Hourlyresidualload for all regionsand storage
12
Finally, after getting the [17 x 8760] LF matrix, several first indicators in Table 10 tell us how
much is expected to be the capacity required per line, the storage to-from lines are omitted.
These are the average flow per line, which we can use to optimize and select the optimal capacity
requirements in each line through methods such as the residual load duration curves, next
section‟s topic.
Line
1
2
3
4
5
6
7
8
9
10
11
1-2
1-5
1-4
2-3
2-6
2-5
2-4
3-6
3-5
6-5
5-4
Max [MW]
9815
7703
2405
9065
4990
3261
2002
420
2025
1913
512
Mean[MW]
1624
2367
734
3362
2144
1422
3
-89
-1362
-1270
-1840
Table 10: LF Indicators for each Line
6.2.2. Duration Curves of the Transmission Performance (DCTP)
A DCTP or the RLDC term used in this context shows the relationship between the load flow
capacity requirements and its utilization period. It becomes thus possible to eliminate the
unnecessarily huge load flow requirements which take place for a limited duration of time.
52
6. Analysis of Germany Divided into Regions
The load flow is first ordered in descending order of magnitude, and at those points of curve
infliction, the limitation and boundary values are taken.
Figure 38 shows the DCTP of line 1 (connecting region 1 to region 2), and the histogram in
Figure 39 is also used, graphically displaying the DCTP and showing the frequency a certain
capacity has been used.
Residual Load Duration Curve for Line 1 (1-2) (Absolute values)
16000
14000
X: 928
Y: 1.2e+004
Residual Load MW
12000
10000
X: 6477
Y: 6570
8000
6000
4000
2000
0
0
1000
2000
3000
4000
5000
Time [Hours]
6000
7000
8000
9000
Figure 38: DCTP for Line 1 with Absolute Values
With the arranged absolute values of the RL flow in line 1 used, we get that:
The TP is always between 0 MW and ≈15300 MW.
The TP is greater than 12000 MW for 1000 hours/year (≈11% of the time)
The TP is between ≈ 6100 and 12000 MW for 5500 hours/year (≈63% of the time)
This is useful in giving us a sense of the utilization period of any capacity that we want to study,
furthermore, and in order to complement the picture, the histogram in Figure 39 show us in an
example for Line1 how much a certain capacity range is utilized and what it is excluded.
Line 1 - RL Histogram
800
700
Times used
600
Exclusion
Region:
capacity
over
6892 MW
not
included
500
400
300
200
100
0
-4000
-2000
0
2000
4000
6000
Residual Load (MW)
8000
10000
Figure 39: Histogram for the Frequency of Use of Line 1 Capacity
53
6. Analysis of Germany Divided into Regions
It can be seen from the histogram and Table 11 that for line 1 and capacity over 6892 MW, the
utilization time is 57 hours of the year (≈0.65% yearly). Which indicates that other transmission
lines (detours in place of the direct path),or other measures like load management methods, at
these hours, should be used to supply the lacking loads.
A refined result is thus reached by applying this concept on each line and setting a criteria for
elemenating capacity at specified frequencies. First, all the possibilities for maximum capacity
reduction are studied for all lines, and then an elimenation percentage for each line is set. In
general we can see a certain average elimination percentage for all lines less than ≈ 1.4%
Finally, it must be noted, that for storage lines no refining or reduction in maximum flows are
made, neither inclusion of their respective capacity because the lines connected to the storage are
not used. They serve as an indication of the generation or consumption needs in each region.
Line
1
2
3
5
6
7
8
10
11
13
11
1-2
1-5
1-4
2-3
2-6
2-5
2-4
3-6
3-5
6-5
5-4
Frequency Percentage Capacity
of use
eliminated new max
over new
or min
maximum
limit
capacity
[MW]
57
0,65%
6892
59
0,67%
5903
115
1,31%
1695
54
0,62%
7828
83
0,95%
4255
2
0,02%
2729
80
0,91%
3359
14
0,16%
882
66
0,75%
4379
74
0,84%
3372
74
0,84%
-6984
Max
capacity
[MW]
Min
capacity
[MW]
Capacity
reduction
percentage
[%]
9815
7703
2405
9065
4990
3261
2002
420
2025
1913
512
-2607
-1298
-1146
-3306
-2355
-2057
-4699
-1027
-5091
-3959
-7799
30%
23%
30%
14%
15%
16%
29%
14%
14%
15%
10%
Table 11: Optimization of Load Flow Capacity Requirements
It is necessary to recall that two storage models were suggested; the bypass (or large distance)
storage model and the central storage model. In Appendix B the different results are shown for
the central storage. We note that LF numbers are less for the lines passing from the regions to
one another, meaning that the rest was exchanged with the storage. This proves our assumption
that more capacity will pass through the storage node if it was put in the middle and that the
central storage model should not be used for demonstration purposes.
54
6. Analysis of Germany Divided into Regions
Regional DC Power Flow
55
14000
54
3
6
53
12000
Latitude
5
51
2
8000
50
MW
10000
52
6000
49
1
4000
4
48
2000
47
4
6
8
10
12
14
16
Longitude
Figure 40: DC Load Flow in Germany with Bypass Storage18
18
The colored arrows are indicative only of the flow capacity in Table 11; the arrows on the circumference of the DC network are the storage‟s generation peaks.
55
6. Analysis of Germany Divided into Regions
6.2.3. Discussions and Conclusions of the Results of the Transport Capacity
The situation regarding the coverage of the residual load and the ability of the different regions
to be an energy importer or exporter of energy is concluded from Figures 36 and 37.
From these figures we see that the north-east, north-west and mid-east regions will act as energy
exporters in the future.
The effect of the prevailing source can be clearly seen; with high wind capacity in the northern
regions the residual load was brought to a minimum, while in these regions with low wind
contribution, the residual load was on average much greater than zero.
A DCLFM was used to get the values for load flows in each line connecting the regions together
in a gird. In order to see how much the fluctuation in storage is, two models were necessary to
simulate for the storage, one which was central and one placed in a distant location. The large
distance storage was used since results showed a greater summation of load flow exchanged
between the regions before going to the storage. This value was 44,111 MW for the non –storage
connected lines, for the large distance model and 29,362 MW for the central storage model
which justifies using the large distance model.
A DCTP was used see the utilization period for any line and then to eliminate maximum capacity
that are utilized for a low number of hours and for high capacity. In general the capacity
eliminated was for utilization periods less than 1.4%, and this had a great effect on reducing
maximum capacity requirements which were 59,813 MW in total and reduced to 48,278 MW, a
reduction of ≈ 19%
56
7. Comparison with the Reference Scenario and Conclusions
7. Comparison with the Reference Scenario and Conclusions
A future reference scenario for Germany is presented, where installed PV capacity is given, for
open spaces along railways, highways, and for roofs and facades. The overall capacity for the
reference scenario was adapted to the six regions scenario. The allocation of the capacity shows
an expected future situation, if the general conditions for the expansion of PV stay the same as
today.
This reference scenario is compared with the scenarios presented earlier, namely: Germany
divided into regions from chapter six and regions that were extracted from Germany as one
region from chapter five.
7.1. Regional Scenario versus Reference Scenario
The reference scenario re-distributes the installed PV capacity in the six regions, based on the
situation and conditions for PV in Germany today.
Region 3
Region 6
55
100
0
54
-50
53.5
-100
53
-150
0
-100
-200
-300
-400
-500
-600
-200
52.5
Difference [MW]
50
Difference [MW]
54.5
Region 2
Region 5
200
0
51.5
-200
51
50.5
-400
50
-600
100
Difference [MW]
52
0
-100
-200
Difference [MW]
52.5
-300
Region 1
Region 4
50
100
49
0
48.5
-100
48
-200
47.5
47
6
7
8
9
10
11
-300
10.5
0
-200
-400
11
11.5
12
12.5
13
13.5
Difference [MW]
49.5
Difference [MW]
200
14
Figure 41: PV Generation Difference between the Installed Regional Capacity and the Reference Scenario
57
7. Comparison with the Reference Scenario and Conclusions
We see in Figure 41 that generally for each region new capacity was added to pixels that had no
potential accounted for before, and that there is a decrease in the potential used for the reference
scenario from what is used in the six regions scenario. Region 6 had particularly new potential
added for in most pixels and opposed to region 1 that had more potential reduced in many pixels.
Since all wind potential was installed for all regions19, there was no need to compare the total
capacity installed in each region with the other two scenarios.
The following table describes the differences between the installed PV capacity in this reference
scenario and the regional scenario.
Ref. [GW]
Sim. [GW]
Match [Ref./Sim.]
Ref. PV generation
[TWh/year]
Sim. PV generation
[TWh/year]
Match [Ref./Sim.]
Ref. PV:wind
Power [%]
Sim. PV:Wind
Power [%]
Region 1
29.12
29.19
99.8%
27.34
Region 2
45.06
45.88
98.2%
39.02
Region 3
17.54
11.96
146.7%
15.93
23.38
40.51
116.9%
85.95:14.15
96.32%
67.96:32.03
143.5%
109%
81.6%
108.1%
37.87:62.13 96.09:3.91 53.39:46.61 40.88:59.12
85.98:14.02
68.36:31.64
29.36:70.64 96.02:3.98 58.90:41.00 39.30:60.70
11.10
Region 4
20.66
20.29
102%
19.25
17.66
Region 5
23.1
28.90
80%
20.71
Region 6
12.14
11.37
107%
11.31
25.38
10.46
Table 12: Differences in Installed PV Capacity between the Reference Scenario and Simulation Results for
the Six Regions
We generally see the good matching between the PV potential used and the newly provided PV
potential from the reference scenario. This is particularly true for regions 1, 2, 4 and 6. The last
two scenarios show the PV to wind share of installed power per region for both scenarios.
Next, we shall investigate the reference scenario‟s effect on the line capacity requirements.
We see in Table 13 that the maximum flow requirements are 12.33% more for the reference
scenario but with almost no change to the mean flow requirements between the two scenarios.
We also see that the flow direction is the same for all maximum flows but is reversed for the
mean flow through lines 8, 9, and 10.
The difference in the maximum capacity requirements is calculated by subtraction of the
maximum flow of the reference scenario from the regional scenario showing that the value for
the reference scenario is higher by 6208 MW.
19
Except for region 6where it is 96.7%
58
7. Comparison with the Reference Scenario and Conclusions
Line
1
2
3
4
5
6
7
8
9
10
11
Max (Ref./Sim. )
[MW]
1-2
1-5
1-4
2-3
2-6
2-5
2-4
3-6
3-5
6-5
5-4
SUM
Percentage
Difference
(Sim. – Ref.)
±[ MW]
6823
9815
-215
6899
7703
960
2059
2405
-383
10586
9065
-1755
6420
4990
-1558
4833
3261
-1691
3264
2002
-2694
2204
420
-1823
2672
2025
-1521
2415
1913
-1704
2146
512
-4279
50320
44112
-6208
-12.33% (Decreased Max. Flow)
Mean (Ref./
Sim. ) [MW]
Difference
±[ MW]
1357
1624
267
2195
2362
167
678
734
56
3595
3362
-233
2208
2145
-63
1455
1422
-33
99
3
-96
152
-89
-241
1514 -1362
-2876
1317 -1270
2587
1720 -1840
3560
16290 16213
-77
-0.47% (Decreased mean flow)
Table 13: Differences in Installed PV Capacity between the Reference and Six Regions Scenarios
Finally, the theoretical accumulated storage was calculated for both scenarios and was found to
be 141.737 TWh for the regional scenario and 144.374 TWh for the reference scenario
increasing the storage requirements.
7.2. Regional Scenario versus Germany as One Region Scenario
Table 14 compares between the installed capacity in Germany as one region scenario and the
regional scenario.
German scenario[GW]
Simulation [GW]
Match [Ger./Sim.]
Yearly Ger. PV Generation
[TWh/year]20
Yearly Sim. PV Generation
[TWh/year]
Match [Ger./Sim.]
German Scenario PV:wind Mix [%]
Regional Scenario
PV:Wind Mix [%]
Region 1
20.83
29.19
71.36%
16.31
Region 2
65.64
45.88
143%
51.38
Region 3
9.17
11.96
76.63%
7.17
Region 4
27.44
20.29
135.25%
21.48
Region 5
19.39
28.90
67.1%
15.180
Region 6
0.078
11.37
0.68%
0.061
23.38
40.51
11.10
17.66
25.38
10.46
69.76%
85.5:14.5
126.8%
75.5:24.5
64.6%
24.2:75.8
121.63%
97:3
59.81%
49:51
81.4:18.6
61.5:38.5
22.5:77.5
96:4
59:41
0.583%
0.44:99.5
6
39:61
Table 14: Differences in installed PV Capacity between Germany as One Region Scenario and Simulation
Results for the Six Regions
20
The value for each region in this row was calculated by dividing the percentage of the PV plants installed in GW
from the total installed by the total sum of the PV generated power per year for Germany unlike Table 12 for the
reference scenario where it was extracted directly from the given input data.
59
7. Comparison with the Reference Scenario and Conclusions
A bigger difference is clearly seen between the capacity installed in Germany as one region
scenario and the regional scenario, this is obvious for region 6 (north-east) where in the one
region scenario almost no PV capacity is installed in comparison with the regional scenario with
99.4% less capacity installed in the Germany as one region scenario.
As in the first comparison, we compare in Table 15 between the differences affecting the
maximum and mean load flow in each of the two scenarios investigated.
Line
1
2
3
4
5
6
7
8
9
10
11
Max (Ger./Sim. )
[MW]
1-2
1-5
1-4
2-3
2-6
2-5
2-4
3-6
3-5
6-5
5-4
SUM
Percentage
Difference
(Sim. – Ger.)
±[ MW]
10030
9815
-215
6743
7703
960
2788
2405
-383
10820
9065
-1755
6548
4990
-1558
4952
3261
-1691
4696
2002
-2694
2243
420
-1823
3546
2025
-1521
3618
1913
-1704
4792
512
-4279
60775
44112
-16663
-27,42% (Decreased Max. flow)
Mean (Ger./
Sim. ) [MW]
Difference
±[ MW]
1838
1624
-214
2286
2362
75
937
734
-203
3235
3362
128
1770
2145
374
1144
1422
278
162
3
-159
-280
-89
191
-1509 -1362
147
-1077 -1270
-194
-1210 -1840
-630
15449 16213
764
+4.71% (Increased mean flow)
Table 15: Differences in Installed PV Capacity between Germany as One Region Scenario and Simulation
Results for Six Regions
Note in Table 15, that there has been no sign change for the lines direction already introduced in
the previous chapter, but that the direction has changed for the mean flows through lines 8,9,10
and 11, for both scenarios.
The sum value shows the sum of the maximum and the mean load flow capacity for all lines, for
each the reference scenario and the simulation scenario. The difference between the simulation
result and the reference scenario is calculated. We see a clear difference of -16,663 MW for the
maximum capacity and 764 MW increase for the mean value. Finally, the percentages show this
reduction for the maximum capacity, and increase for the mean value.
Finally, the theoretical accumulated storage was calculated again these two scenarios and was
found to be 141.737 TWh for the regional scenario and 160.850 TWh for the reference scenario
increasing by almost 13% the storage requirements.
60
7. Comparison with the Reference Scenario and Conclusions
7.3. Conclusions of the Comparison
The redistribution of PV capacity in the reference scenario affected matching between the two
scenarios with regards to the capacity installed, and energy generated yearly.
Also, maximum flow capacity for all lines increased by 12.33% with the reference scenario,
meaning that the regional scenario has less capacity requirements for lines and thus is a better
scenario for load flow since the requirements for capacity are less. There was almost no change
on the mean flow for both scenarios and thus no preference can be based on this aspect.
Taking the comparison with Germany as one region scenario, we see that the overall capacity
installed is more in the regional model, which when summed up is 147.59 GW for the six regions
scenario and is 142.54 GW for the Germany as one region scenario.
When we redistribute capacity for Germany as one region to match the regional model, load flow
requirements increase by a significant 27% for the maximum flow and reduced by almost 5% for
the mean flow. And so, it is even clearer in comparing with this scenario that less flow capacity
is needed with the Germany as one region division to regions than having a Germany as one
region scenario. Accordingly, it is useful to study regions and have it as a focus point to study
further regional arrangements.
If we compare between results of the reference scenario to the Germany as one region scenario,
we see that the reference scenario installs more PV in the regions providing less flow
requirements, this indicates the better distribution of PV capacity in the reference scenario in
comparison with the Germany as one region scenario.
When taking the storage requirements we can see the advantage of the regional scenario as it
reduces by around 2% and 13% the storage requirements in the reference scenario and the
Germany as one region scenario consecutively.
61
7. Comparison with the Reference Scenario and Conclusions
Finally, figures for the conclusions made thus far are summarized in Table 16.
Total PV Potential
[GW]
PV Potential
Reg./Other scenario
Total PV generation
[TWh/year]
PV:Wind in annual
generation %
LF/Reg.
% Load covered
Storage Size
[TWh/year]
% Storage
requirements/Reg.
scenario
Regional scenario
Reference Scenario
Germany as one region
scenario
147.95
147.62
142.54
- 0.22%
- 3.66%
128.5
133.56
111.58
40,39%
41,63%
37,33%
69.18%
+ 12.33%
69.79%
+ 27.42%
65.01%
141.736934
144.374350
160.850210
+ 1.86%
+ 13.48%
Table 16: Summary of Main Conclusions between the Regional Scenario and the Reference Scenario as well
as Between the Regional Scenario and Germany as One Region Scenario
We notice that this has an effect on the total load demand coverage to be 69.79% for the
reference scenario, slightly more than the 69.18% for the regional scenario because for the
reference scenario the orientation of PV panels is optimized to generate the maximum yield
without giving consideration to what‟s best for the whole supply structure.
62
8. Case study: Morocco
8. Case study: Morocco
Morocco has a high potential in both, wind with a 1188 TWh/year technical potential, and solar
power with a Global Horizontal Irradiance (GHI) for PV of 2000 kWh/m2/year [11].
The purpose of this chapter is to demonstrate the rough figures on reducing the average residual
load to a zero by installing PV and wind capacity through the allocation methods.
Due to the lack of available detailed data, in contrary to the analysis for Germany, 60 locations
were chosen in the country where 60 solar and 51 wind data normalized generation time-series
were extracted, and in so doing, the study reduced to only common 51 points. This meant that
independent scenarios for PV and wind are looked-for. This is shown in Figure 42.
PV and Wind data selection for Morroco
36
Wind data (concentrated)
PV data (dispersed)
35
56
57
58
59
60
50
51
52
53
54
55
34
42
43
44
45
46
47
48
49
33
34
35
36
37
38
39
40
41
25
26
27
28
29
30
31
32
18
19
20
21
22
23
24
12
13
14
15
16
17
Latitude
33
32
31
X: -9.786
Y: 29.81
30
29
7
8
9
10
2
3
4
5
28
1
27
-14
-12
-10
11
X: -8.985
Y: 27.73
6
-8
-6
-4
-2
0
Longitude
Figure 42: Selected Points for PV (Black-Dispersed) and Wind (Green-Concentrated)
63
8. Case study: Morocco
8.1. Data
8.1.1. Solar Data
The solar data for the year 2005 was extracted per pixel from the SODA (Solar Energy Services
for Professionals) data base on their website for 60 locations in Morocco [12]
The Solar data for Morocco extracted from the SoDa website were: the hourly mean clear sky
model and the global horizontal irradiance as well as the ambient temperatures.
The data was processed and normalized time-series for the different tilt and orientation angles
developed with the Matlab program [13].
8.1.2. Wind data [14]
Wind data that covers the range of [7km x 7km] area was extracted from the COSMO-EU data,
after that, this large set of data was calculated near the chosen coordinates of the PV since it
determined the number of locations to be studied.
The methodology is based on a physical model with modeled wind fields from a numerical
weather prediction model as a starting point. It includes reductions of wind speed and power
outputs due to shading effects, grid losses and availability as well as assumptions about future
hub heights, future power curves, the dismantling of older wind turbines and their replacement
by modern turbines and the additional reduction of the availability of offshore wind farms on
stormy days.
For this particular case, average wind speeds at the common PV-Wind points were taken, and the
power curve at each point generated, normalized power time-series for these points, Figure 43
show the power curve at the coordinate point (Long. -9,17 Lat. 28,66).
Power Curve
1
0.9
normalised power
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
5
10
15
wind speed [m/s]
20
25
Figure 43: Power Curves at a Specified Location in Morocco
64
8. Case study: Morocco
The power curve was made by a convolution of the original power curve (of a MM92-turbine)
with a Gaussian distribution with a standard deviation of [0, 1 x wind speed]. The convoluted
power curve leads to a smoother shape of the power curve, and in doing this we model the spatial
distribution of the turbines within a pixel. The more turbines we have within a pixel the smoother
the "pixel-power curve" should be.
The hub height used and measurements taken were at an altitude of 100m. The turbine REpower
MM92, and wind speed values were taken every 1 hour. The power output was reduced by 5%
due to electrical losses and turbine failures.
Finally, the nominal power of the MM92-turbine is 2 MW. And the smoothing of the power
curve seen in Figure 44 is the "multi-turbine power curve"[15].
8.1.3. The load curve
As shown in Figure 44, the load curve of Morocco - provided by ONE [16] - is significantly less
than that of Germany. Several indicators for this significance are given in Table 17.
8
x 10
7
4
Germany Load Demand 2009
X: 449
Y: 7.446e+004
MW
6
5
4
X: 3629
Y: 2.995e+004
3
2
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
8000
9000
Time[Hours]
Morocco Load Demand 2010
5000
X: 5350
Y: 4745
MW
4000
3000
X: 7685
Y: 1714
2000
1000
0
1000
2000
3000
4000
5000
6000
7000
Time[Hours]
Figure 44: The Load Curve of Germany and Morocco for 2009 and 2010 Respectively
65
8. Case study: Morocco
We see from the Figure 44 and Table 17 the huge difference between the two load curves, and
this would definitely affect the simulation results for Morocco.
Load Peak GW
Load Mean GW
Yearly Load TWh/Year
High Peak [day]
Low peak [day]
Morroco
5.35
3.142
27.523
12.08
19.11
Germany
74.464
52.482
460
19.01 & 21.11
6.03 & 13.04
Table 17: Load Demand Characteristics in Morocco and Germany
8.2. Experimental Framework
The aim of the study was to see the capacity required to cover 100% average residual load by
each source independent of the other. This objective was set as we do not have an equal number
of data points for both PV and wind.
The scenarios developed to demonstrate the results for each source will depend directly upon the
factors used in the suitability factor formula. These factors are shown in Table 15 for both
scenarios.
Taken the area of Morocco, the assumption made for the wind potential in each ≈ 7700 km2 is set
at a safe 1000 MW.
As for PV, with more pixels to study, each location covered a range of ≈7500km2 and was set
also at 1000 MW PV potential per location.
Weighing the three allocation methods for wind produced three scenarios. This allows us to
understand which factors contribute the best in covering the residual load with the least capacity
installed. The third allocation method for PV was omitted from the calculations as will be
explained in section 8.3.2
The following table shows the different scenarios, Table 18.
To the power 3
Correlation coefficient’s power (PowerCC_PV)
Covering the load power (PowerDA)
Covering the residual peak power
(PowerDA_LS)
Wind scenario
W_A (3,1,1)
W_B (1,3,1)
W_C (1,1,3)
PV scenario
PV_A(3,1)
PV_A(1,3)
-
Table 18: Scenarios for the Coverage of the Residual Load of Morocco by PV and Wind
66
8. Case study: Morocco
8.3. Simulation
8.3.1. Wind Scenarios
The results show a consistent installed wind capacity of 12,250 MW that brings the average
residual load to near a zero value, moreover, the hours in which the residual load is less than zero
were around 3,410, or around ≈ 40% of the time for the three methods.
Table 16 shows a summary of the results for the three methods. Figure 45 shows the locations for
the installation of wind capacity for scenario W_A. Figure 46 shows the wind generation versus
load demand.
Installed wind
capacity [MW]
Average residual load
[MW]
Hours where residual
<=0
Peak residual load
reduced
W_A
12250
W_B
12530
W_C
12250
-1,964
-0,604
-1,435
3418
3411
3407
≈ 7%
≈ 7%
≈ 7%
Table 19: Summary of Wind Scenario Results
Installed wind capacity (12250 MW)
1000
36
35
800
34
Latitude
33
600
32
31
400
30
29
200
28
27
-12
-10
-8
-6
Longitude
-4
-2
0
Figure 45: Installed Wind Capacity for Morocco
67
8. Case study: Morocco
It is shown, that the eastern region has good pixels for the installation of wind capacity, but this
doesn‟t tell if there are barriers to the installation of wind farms in those regions, as the real
potential is not known. Also, the north-western region is an adequate location for the installation
of wind capacity.
100% Wind Scenario
9000
Load Demand
Wind Generation
8000
7000
MW
6000
5000
4000
3000
2000
1000
0
4
8
12
16
20
24
28
Time [Week]
32
36
40
44
48
52
Figure 46: 100% Wind Share Scenario for the Coverage of the Residual Load of Morocco
In Figure 46, we see that a 100% wind scenario only, has no coverage over the summer period,
and an excessive coverage between 44-48 weeks. This would require high storage and transfer
capacity.
8.3.2. PV scenarios
If we are to take the generated power from PV alone, we must first realize that the fluctuation of
this source is diurnal, with a trend of increased generation during summer time, and that no
generation is available during night time.
This means, that no matter how much we install PV capacity, it will not serve our objective of
reaching a zero residual load when for times with no PV generation. All the extra capacity
installed would not be required during the load demand hours, unless we have enough storage to
supply this energy at night or transfer mechanism for this energy with neighbouring countries.
The analysis here is only concerned with presenting rough values for the capacity needed to
reach a zero residual load mean, and another figure for how much capacity is needed if no
storage or transfer capacity is available. For that, a step size of 40 MW was chosen.
68
8. Case study: Morocco
Additionally, discussion about the ideal installation angles (tilt and orient) is presented, as well
as the ideal locations for this installation.
Only two of the three allocation methods were used, namely: the correlation coefficient method
and covering the load method. The contribution of the load peaks coverage is not an objective to
investigate in this scenario, as the peak is falling in a time period with no PV generation, which
renders it impossible to reduce.
Figure 47 explains the impossibility of covering loads and load peaks at regions where there is
no PV generation and the inherent need to mix with other renewable sources, storage or transfer.
1
Load demand
PV generation at 0 tilt, 70 orient, pixel 1
No PV generation zone,
no coverage of peaks and load possible
0.9
0.8
0.7
pu
0.6
0.5
0.4
0.3
0.2
0.1
0
0
20
40
60
80
100
120
140
160
Time [Hours]
Figure 47: No PV Generation at Night time and Comparison with the Load
As can be seen, less installed capacity is required with more emphasis on covering the load
method, this is obvious, as the correlation will be somehow constant and rising the term of the
correlation coefficient to the power three will not increase the load covered.
Installed PV capacity
[MW]
Average residual load
[MW]
Hours where residual
<=0
Peak residual load
reduced
PV_A (CC^3)
25560
PV_B (DA_PV^3)
19520
-4,2
-2.5
3842
3722
0%
0%
Table 20: Comparison between the Two PV Results Based on Different Weighing Factors
69
8. Case study: Morocco
Table 20 shows the difference in the installed PV capacity per each method weight. The PV_A
has the correlation coefficient power equal 3 and a load coverage power equal to 1. While the
PV_B scenario has a load coverage power of 3 and a correlation coefficient power equal to 1. In
the coming figures, results for PV_B scenario are displayed.
In Figure 48, we can see the locations of the installed PV capacity, these are primarily in the
south west, and north east regions.
Morroco - Installed PV capacity
36
1000
35
900
800
34
700
33
Latitude
600
32
500
31
400
30
300
29
200
28
27
-14
100
-12
-10
-8
-6
Longitude
-4
-2
0
0
Figure 48: Installed PV Capacity of 19520 that is Able to Bring the Mean Residual Load to Zero
However, more regions in the northern west locations were also suitable according to the
installation methods and factors used.
In Figures 49 and 50, we see the ideal tilt and orientation angles were the highest frequent tilt
and orient angles were 50° and 80° respectively.
70
8. Case study: Morocco
Used tilt and orientation angles for Morroco
100
5000
90
4500
4000
70
60
3500
50
3000
40
2500
30
2000
20
1500
10
1000
0
500
-10
100
80
60
40
20
West
0
-20
-40
South
-60
-80
Times Used
Tilt Angle (Inclination)
80
0
-100
East
Figure 49: Tilt and Orientation angles for the PV scenario
Frequency
0.4
0.3
X: 80
Y: 0.3
X: 50
Y: 0.4
X: 30
Y: 0.2
0.2
0.1
0
0
10
20
30
40
50
60
70
80
90
Tilt
Frequency
0.4
0.3
0.2
X: -90
Y: 0.296
X: 80
Y: 0.278
0.1
0
-100
X: -30
Y: 0.105
-80
-60
-40
-20
0
20
Orientation( W S E
40
60
80
100
)
Figure 50: Frequency of Tilt and Orientation Angles for the PV Scenario
In Figure 51, the weekly average PV generation curve that is able to overcome 100% residual
load by PV only is shown in the figure. Keep in mind that the values for both PV and the load
demand are averaged on weekly basis, and so, we cannot see the no-generation periods of PV on
daily basis.
71
8. Case study: Morocco
We see from the figure that PV generation is highest between the 20th and 30th weeks, which
occur in summer and is lowest in winter. There is a slight correlation with the load which
experiences a peak between the 33rd and 40th weeks.
100% PV scenario
4000
3500
MW
3000
2500
2000
1500
0
5
10
15
20
25
30
35
40
45
50
Time[Week]
Figure 51: PV Generation Curve for a 100% PV Scenario
8.3.3. Storage Requirements
5
20
PV Accumulated Storage
x 10
MW
15
10
5
0
-5
0
1000
2000
3000
6
1
4000
5000
6000
7000
8000
9000
6000
7000
8000
9000
Wind Accumulated Storage
x 10
MW
0
-1
-2
0
1000
2000
3000
4000
5000
Figure 52: Hourly Theoretical Accumulated Storage for PV and Wind
72
8. Case study: Morocco
In Figure 52 the theoretical needed storage size is found out by summing up the storage or
generation needed from the previous hour with the current hour. Two theoretical sizes were
found for the PV and wind models. The size for PV size is 2.008602 TWh and for wind is
2.552547 TWh.
4
4
6
Daily Storage - PV
x 10
1
0.5
0
MW
MW
2
-2
0
50
100
5
2
0
-0.5
-4
-6
Monthly Storage - PV
x 10
150
200
250
300
-1
350
Daily Storage - Wind
x 10
1
2
3
6
1.5
4
5
6
7
8
9
10
11
12
9
10
11
12
Monthly Storage - Wind
x 10
1
MW
MW
1
0
0.5
0
-0.5
-1
0
50
100
150
200
Time[Days]
250
300
350
-1
1
2
3
4
5
6
7
Time[Days]
8
Figure 53: Storage Requirements for 100% PV and Wind Scenarios in Morocco
As we can see in Figure 53, there is a clear difference between the uniformity of storage
requirements for PV and wind, it is clear for PV that during summer months there is an excess of
energy generation and a lack thereof during winter. The maximum daily stored or discharged
energy was calculated and found to be 45,147 MWh/day and a maximum monthly stored or
discharged energy of 803,011 MWh/month. The range between this maximum and minimum
storage or discharge required is 45,108 MWh/day and 710,123 MWh/month respectively.
The sum of stored and discharged energy during the whole year for PV is 14,371,104 MWh/year
and 14,407,975 MWh/year respectively, which leads to the hourly average -2.5 MW residual
load.
For wind, there is less uniform storage or discharge utilization temporally. The maximum daily
stored or discharged energy is found to be 189,022 MWh/day and a maximum monthly stored or
discharged energy of 1,452,732 MWh/month. The range between this maximum and minimum
storage or discharge required is 188,811 MWh/day and 1,438,623 MWh/month respectively.
The sum of stored and discharged energy during the whole year for wind is 10,589,109
MWh/year and 10,606,316 MWh/year respectively, which leads to the hourly average -1.96 MW
residual load.
73
8. Case study: Morocco
Discussion of the results and conclusions
In comparing the three wind scenarios together, no clear conclusions could be made about the
effect of the allocation methods factors. This is mainly due to the rough estimation of wind
potential available at each site, and gross size of the pixels used.
For a wind-only scenario, a capacity of 12,250 MW will be needed to cover the residual load,
however, this doesn‟t take into account the storage needed to supplement the needed energy at
times of no wind generation, i.e. week numbers 7 and 24-36.
The 100% PV scenario was studied and the least capacity needed to cover the residual load is by
using the coverage of the load term to equal 3.This can be understood, as our aim is to cover the
mean residual load as a bulk quantity, whereas the correlation coefficient only looks at the
correlation between PV generation and load at each hour
For the selected PV-only scenario, a capacity of 19,520 MW will be needed. The argument that
this capacity is not well dispersed to match the load curve stands as well, and renders this figure
much higher than the actual needed, because storage will be needed to store the excess capacity
at day time and transfer it at night time to the grid.
From the storage requirements, it is clear that storage in summer is needed for the PV scenario to
supplement it in winter. This cannot be clearly concluded for wind due to the dispersion of
generation throughout the year, although there is a general trend of increased generation when
PV is lacking generation, i.e. during winter months.
The ratio between storage requirements for PV based on the total sum of energy stored or
discharged daily is only 24% of wind daily storage requirements, but when calculated monthly
becomes 55% of the wind monthly storage requirements, finally, these storage requirements
become 136% more for PV yearly storage than yearly wind storage requirements. We conclude
that daily storage for PV is favorable to compensate the day and night gap, and that a large
yearly or seasonal capacity storage for wind would be a better storage solution for the wind
scenario.
We notice from Figure 52 that the accumulated storage requirements for PV are less than for
wind with a max of 1611788 MW for PV and -1695790 for wind
74
9. Conclusions
9. Conclusions and Recommendations for Further Research
9.1. Conclusions
According to current assumptions, the PV potential is higher than wind in Germany; the total
available PV potential in this study is around 380 GW, while it is 95 GW for wind. With further
development of the technique, declining prices and change in regulations, the usable potential for
wind may increase.
Both sources are able to overcome a significant part of the mean residual load and reduce it by
around 65% when installing only half of their available capacity of 237.500 GW with a total
yearly generation of 298.880 TWh/year. However, only 7% of the maximum residual load is
reduced mainly due to the incidence of the load demand peaks at time periods with no suitable
PV or wind power generation.
Installing half the total capacity for both sources resulted in installing the full wind potential and
only 37.45% PV potential.
Most of the installed PV capacity was in the west, south and south-east regions, because there are
more buildings and more highways and railways and thus more potential. As for wind, the northwestern region dominated the installed wind capacity.
The most frequently used tilt angle in the Germany as one region scenario was 70° and were
60°, -10°and -60° for orient angle. This deviates from the situation in place nowadays in
Germany, were PV tilt angles are frequently set at 32.5° and the orient angle on the other hand is
set at 10°. The main reason for this deviation is that energy producers put more emphasis on the
bulk energy generated by PV modules than other considerations accounted for in this study,
which looks at the supply structure, its correlation with the load demand, and the distribution of
the load coverage daily and seasonally, this eventually reduces storage and transmission
capacity.
When Germany is divided into regions, it becomes possible to see how the new regions reduce
the load demand, increasing the coverage of the mean residual load to 69.18%. The total yearly
generation was found to be 318.049 TWh/year, 6.4% more than the Germany as one region
model.
According to the definition of the regions, the south-west, mid-west and south-east regions were
able to reduce the mean residual load by 31%, 54% and 37% respectively. The main source
installed in these regions was PV. The angles installed in each region also differed than what is
75
9. Conclusions
in place nowadays in Germany. However, different regions exhibit different “ideal” tilt and
orientation angles.
Wind energy is the leading source installed in the north-east and north-west regions, it is also
considered to be the leading source in the mid-east region. In these regions, as the situation in
Germany as one region, all the available wind potential was installed.
With almost 139% reduction in the mean residual load, the north-west region will be an energy
exporter in a 100% PV-wind energy scenario. The north-east and mid-east regions will also play
this role. This can be read in combination with the low load demand due to the less population
density, particularly in the north-west and north-east regions.
Transport and storage capacities are possible to model for the regional scenario. Based on the
storage duration curve, we can see that during 25% of the time, energy is being exported from
the regions while it is being transferred to the regions for the other 75% of the time. A grid
connecting the centers of the regions was created using the DCLFM to export and import energy
first between the regions before all the excess or lack energy gets exported or transported to the
ideal storage/generator. From the load flow in each line, the range of line capacity is between
800 MW and 7800 MW; the minimum and maximum values of load flow through the DC grid.
Two scenarios were used to compare the load flow requirements and PV capacity installed in the
regional scenario to provide concrete answers to whether our objective is achieved of predicting
qualitative indicators of load coverage and PV-wind installation by the regional model.
In both cases, the maximum flow requirements were less in the regional model by 12% and 27%
than for the first reference scenario and the second Germany as one region scenario. The storage
size was found to be 160.850 TWh and 144.374 TWh for the Germany as one region scenario
and the reference scenario respectively. This is higher by 13% and 2% than the 141.737 TWh
storage capacity calculated for the regional model. These results for the storage and transmission
requirements have proved the advantage of using the regional model analysis to reduce both the
storage and transmission requirements.
The PV share from the total installed capacity was 41.63%, 40.39% and 37.33% for the reference
scenario, regional scenario and Germany as one region scenario respectively. The total load
coverage was most for the reference scenario because of the higher energy yield with a value of
69.79% slightly higher than the regional scenario with a value of 69.18%. The total load
coverage for Germany as one region scenario is found to be 65.01% owing to the fact that less
PV was installed.
Based on these results, we conclude that further divisions in the map of Germany is
advantageous to study the ideal positioning of PV and wind capacity as well as further reducing
transport and storage capacity as well as the mean residual load.
76
9. Conclusions
In Morocco case study, 1000 MW for each PV and wind source potential was able to overcome
the mean residual load by a 100% PV scenario and by another 100% wind scenario. For wind,
12,250 MW were installed to meet the mean residual load. For PV, the installed capacity value
was 20 GW when a bigger weight was given for covering the load factor and was 25 GW when
the weight was bigger for the correlation with the load factor. The most frequent PV tilt and
orient angles for the 20 GW capacity are 50°, 80° for tilt and 90°, -90° for orient angles.
Although less capacity was needed to overcome the mean load by wind, more storage capacity is
needed and was found to be 2.553 TWh, exceeding by 27% the requirement for PV storage of
2.009 TWh.
9.2. Recommendations for Further Research
 Investigating the temporal impact by using input from more than one year
 More renewable energy sources must be included, this study only focused on the
fluctuating sources PV and wind, however, other sources such as Bio-energy and
Geothermal energy could play a role in providing a coverage of the base load and
contribute to better assessment of load demand coverage possibilities and reducing the
required storage and transportation capacity.
 More regions, and a new configuration and division of these regions, must be studied in
order to approximate the reality and get a better estimation of the loads in each new
region, the available capacity to cover this load, and to ideally place PV, wind and other
RES capacity.
 The weighing of the suitability factor formula terms must be optimized and studied for
each RES in the study, so that the objectives of integrating each RES are properly
fulfilled.
 Considering further criteria for the suitability factor, e.g. transmission capacity and
gradients of the residual load.
 Study new and various DC connections between the regions, and study in details the
losses incurred, and better estimation of the capacity limits in each line.
 Study the real technical requirements for storage and find the best locations for them.
77
References
References
[1] Widen, J. , “Correlations Between Large-Scale Solar and Wind Power in a Future Scenario
for Sweden”, Sustainable Energy, IEEE Transactions on , vol.2, no.2, pp.177-184, April 2011
doi: 10.1109/TSTE.2010.2101620,[Online], Available:
http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5685581&isnumber=5735622
[2] Dominik Heide ,Lueder von Bremen, Martin Greiner, Clemens Hoffmann, Markus
Speckmann, Stefan Bofinger, “ Seasonal optimal mix of wind and solar power in a future,
highly renewable Europe ”,[Online], Available: http://www.iset.uni-kassel.de/abt/FBI/publication/2010-014_Seasonal_optimal_mix.pdf
[3] Thomas Nikolakakis, Vasilis Fthenakis, “The optimum mix of electricity from wind- and
solar-sources in conventional power systems: Evaluating the case for New York State”, [Online],
Available: http://www.sciencedirect.com/science/article/pii/S0301421511004526
[4] Ghassan Zubi, “Technology mix alternatives with high shares of wind power and
photovoltaics—case study for Spain”, Energy Policy, Volume 39, Issue 12, December 2011,
Pages 8070-8077, ISSN 0301-4215, 10.1016/j.enpol.2011.09.068 ,[Online], Available:
http://www.sciencedirect.com/science/article/pii/S0301421511007828
[5] REN 21, “Renewables 2011 - Global Status Report”, [Online], Available:
www.ren21.net/Portals/97/documents/GSR/REN21_GSR2011.pdf
[6] Yi Li; Agelidis, V.G.; Shrivastava, Y.; “Wind-solar resource complementarity and its
combined correlation with electricity load demand” Industrial Electronics and Applications,
2009. ICIEA 2009. 4th IEEE Conference on, vol., no., pp.3623-3628, 25-27 May 2009
doi: 10.1109/ICIEA.2009.5138882, [Online], Available:
http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5138882&isnumber=5138149
[7] Fraunhofer-Institute (IWES), “Vorstudie: Statistische Auswertung der Häufigkeit des
zeilichen Zusammenfallens hoher Einspeisung aus Photovoltaik und Windenergie Statistische
Auswertung der Häufigkeit des zeilichen Zusammenfallens hoher Einspeisung aus Photovoltaik
und Wind”, Page 10, 2.1.1
[8] Fraunhofer-Institute, Kassel-Germany, Contact person: M.Sc. Raphael Spiekermann,
[9] German remote sensing data centre, Info. Retrieved on: 10.Feb.2012
http://www.corine.dfd.dlr.de/intro_de.html
[10] Fraunhofer-Institute, Kassel-Germany, Contact person: Mareike Jentsch
[11] DLR, “WP 3: Renewable Energy Resources in EU-MENA”, [Online], Available:
http://www.dlr.de/tt/Portaldata/41/Resources/dokumente/institut/system/projects/WP3_Resource
s_Final.pdf
78
References
[12] SoDa, Solar Energy Services for Professionals, http://www.sodais.com/eng/services/service_invoke/ gui.php?xml_descript=hc3v2_invoke _15_demo.xml#parameters
[13] Fraunhofer-Institute, Kassel-Germany, Contact person: Kaspar Knorr
[14] Knorr, B. Lange, D. Callies, M. Wolff Institut für Solare Energieversorgungstechnik
e.V., Kassel, Germany, “SIMULATION OF FUTURE TIME SERIES OF WIND POWER
PRODUCTION IN GERMANY BY 2020”
[15] Per Nørgaard, RISØ National Laboratory, Denmark ([email protected])
Hannele Holttinen, VTT, Finland ([email protected]) , “A Multi-Turbine Power Curve
Approach”, NORDIC WIND POWER CONFERNCE, 1-2 MARCH 2004, CHALMERS
UNIVERSITY OF TECHNOLOGY 1
[16] ONE, l'Office National de l'Electricité du Maroc
[17]Wiki answers homepage, [Online], retrieved on 11.Feb.2012
http://wiki.answers.com/Q/What_is_dc_load_flow#ixzz1mf7JKLto
[18] Iowa State University homepage, [Online], retrieved on 11.Feb.2012
http://class.ece.iastate.edu/ee458/PowerFlowEquations.doc
79
Appendices
Appendices
Appendix A: DC Load Flow
DC load flow is a method to estimate power flows through lines on AC power systems. The
following is a brief introduction to the main concept of the power flow used in this study.
A "DC" load flow uses a simplified, linear form of modelling the AC system. Consequently its
solution is non-iterative, and absolutely convergent. It becomes a routine algebra problem,
solving multiple equations with multiple variables. It is inherently less accurate than a "full" AC
load flow solution, but it is useful where fast, dependable solutions are essential, and the
approximation is acceptable.
In reality, there is nothing "DC" about a DC load flow. It solves for phase angles (an AC,
reactive characteristic); it ignores resistance (a DC characteristic); and it ignores voltage
(because the objective is just power flow). It probably derives its name from the similarity
between this solution method and the method used to solve a DC system, which is also linear,
non-iterative and absolutely convergent. [17]
The approximated DC power flow equation is shown Eq. (A.1) [18]
N
Pk
Bkj (
k
j
)
j 1,
j k
A.1
Where, Pk is the real power flow through the network. k , j : are two voltage phasor angles at
buses j and k. And Bkj is the imaginary part of the admittance matrix formula.
Given the total number of buses and lines, the topology of the network, and power flow
injections at all buses except at the slack node bus, the power flow is possible to calculate, and an
example for a 100 MW flow from point A to B through line C is shown in the following figure:
X
300
-500MW
Y
+500MW
200
200
Z
Figure 54: Power flow from X to Y through Z
80
Appendices
Appendix B: Central storage model for the six regions
Figure 55 shows the RLDC for line 1 for the central storage. The storage maximum and
minimum capacity requirements are less than the fictional large distance storage scenario.
Residual Load duration curve for Line 1 (1-2)
7000
X: 1
Y: 6146
6000
Residual Load [MW]
5000
4000
X: 594
Y: 3000
3000
2000
X: 7490
Y: 485.3
1000
0
-1000
-2000
X: 8760
Y: -1787
0
1000
2000
3000
4000
5000
Time [Hours]
6000
7000
8000
9000
Figure 55: RLDC for line 1 (Connecting Centers of Regions 1-2)
Table 21 shows the capacity through lines 1 to 11 for the central storage model.
Line
Repetition
1
2
3
4
5
6
7
8
9
10
11
45
23
17
44
52
36
15
18
83
0
32
Capacity
New max
Percentage
or min
eliminated
limit
[MW]
0,51%
4560
0,26%
4177
0,19%
1545
0,50%
5819
0,59%
3056
0,41%
1988
0,17%
-2201
0,21%
-731
0,95%
-3391
0,00%
0
0,37%
-4422
Max
Min
% Max
reduction
of
capacity
6146
4758
1818
6622
3515
2394
1237
311
1162
754
645
-1787
-1060
-905
-1410
-1082
-1668
-2583
-847
-3897
-2789
-4985
26%
12%
15%
12%
13%
17%
15%
14%
13%
0%
11%
Table 21: Line Parameters for the Central Storage Model
81
Appendices
Appendix C: Installed Wind and PV Capacity for Regions
Installed Wind capacity for Region One (South-West)
50
Installed Wind capacity for Region Four (South-East)
Installed Wind capacity for Region Two (Mid-West)
140
40
50
52.5
400
35
120
52
30
300
60
51.5
250
200
51
49
20
48.5
15
150
48
25
48
40
47.5
47
47.5
20
6
7
8
9
10
11
10
100
50.5
5
50
0
50
5
6
Longitude
7
8
9
10
11
47
10
0
0
11
14
Installed PV capacity for Region Six (North-East)
Installed PV capacity for Region Five (Mid-East)
52.5
55
13
Longitude
Longitude
Installed PV capacity for Region Three (North-West)
12
55
450
300
800
700
400
150
53.5
350
51.5
250
200
51
100
53
50
600
300
MW
Latitude
200
54
54.5
52
250
MW
Latitude
Latitude
54.5
500
54
400
53.5
300
150
100
50.5
200
53
100
50
52.5
6
7
8
9
Longitude
10
11
MW
48.5
MW
Latitude
Latitude
49
MW
100
80
49.5
350
MW
Latitude
49.5
50
10
11
12
13
Longitude
14
15
16
52.5
10
11
12
13
14
15
Longitude
Figure 56: Installed PV and Wind Capacity for All regions where Source is not Leading in Installed Capacity
82
Potenzial einer systematischen Standortwahl bei einem starken Ausbau der
Kapazitäten von PV und Windkraftanlagen in Deutschland
Analyse des Speicher- und Transportbedarfs und eine Fallstudie für Marokko
Eine Masterarbeit im Rahmen des Studiengangs REMENA - eingereicht von Arabi Abdelhaq
Die Welt steht vor neuen Herausforderungen: schwindende Ressourcen, expandierende
Volkswirtschaften und eine stetig wachsende Bevölkerung mit einem ressourcenintensivem
Lebensstil. Folglich müssen Politiker alle verschiedenen Kombinationen des Einsatzes von
erneuerbarer Energien analysieren und ihre Umsetzbarkeit prüfen. Das oberste Ziel sollte die
Gestaltung verschiedener Optionen für eine effektive Nutzung erneuerbarer Energien - jetzt
und in der nahen Zukunft - sein.
Diese vorliegende MATLAB-basierte Studie analysiert das Potential einer systematischoptimierten Standortwahl von Windturbinen und Photovoltaikanlagen (PV) in Deutschland,
die die Residuallast und die verbleibenden Anforderungen an die zukünftige
Energieversorgung minimiert. Zusätzlich wird eine vereinfachte Fallstudie für Marokko
vorgestellt.
Die Studie untersucht die notwendige installierte Kapazität an Photovoltaik- und
Windenergieanlagen, die aus der angewendeten Allokationsmethode resultiert und die
Residuallast in Deutschland signifikant reduziert. Um die Auswirkungen einer optimierten
Verteilung von Anlagen in kleineren Regionen auf das Gesamtsystem deutlich zu machen,
wurde Deutschland in sechs Regionen unterteilt. Dies ermöglicht zudem die Betrachtung der
benötigten Transport und Speicherkapazitäten, die den maximalen Lastanforderungen
genügen.
Das in dieser Studie vorgeschlagene „sechs-Regionen-Szenario“ weist eine installierte
Gesamtkapazität von 238 GW auf und reduziert die Residuallast um 69.18%. Dies stellt eine
Verbesserung im Vergleich zu dem Szenario mit nur einer Region dar, welches eine
Residuallast von 65% aufweist. Der durchschnittliche Lastfluss in den GleichstromÜbertragungsleitungen zwischen den Regionen beträgt 1.474 MW, welche primär für den
Lastausgleich genutz werden, bevor sekundär die Speicher im Bereich von 17 GW und 141.7
TWh Kapazität zum Einsatz kommen.
Die Resultate für die Speicherkapazität und den Transport werden mit zwei Szenarios
verglichen: Deutschland als eine Region und einem zweiten Referenzszenario. Die Ergebnisse
zeigen, dass die notwendige Transportkapazität um 27% und 12% und die Speicherkapazität
um 13% und 2% geringer als in den beiden Vergleichsszenarios ist. Dies unterstreicht die
Bedeutung der Nutzung eines regionalen Optimierungsansatzes, um den Bedarf an Speicherund Übertragungskapazitäten zu reduzieren.
In der Fallstudie für Morocco wird die notwendige PV und Windenergiekapazität, die eine
rein rechnerische Lastabdeckung garantiert, auf jeweils 19.520 MW und 12.250 MW
geschätzt. Die Speicheranforderungen wurden für beide Fälle untersucht und weisen klare
Unterschiede auf: Im Fall der PV wird 27% weniger akkumulierte Speicherkapazität benötigt
als bei der Nutzung von Windenergie.
Kassel, 28.02.12
‫االمكانٌة المتاحة فً توزٌع مزٌج ذي سعة عالٌة من مصادر الرٌاح و الطاقة الشمسٌة بطرٌقة‬
‫منهجٌة فً ألمانٌا‬
‫تحلٌل متطلبات النقل و التخزٌن و دراسة حالة فً المغرب‬
‫رسالة ماجستٌرمقدمة لبرنامج الطاقة الجدٌدة والمتجددة وكفاءة استخدام الطاقة ‪REMENA -‬‬
‫تقدٌم‪ :‬عربً عبد الحق‬
‫الملخص التنفٌذي‬
‫ٌواجه العالم تحدٌات جدٌدة تتعلق باستنزاف الموارد الطبٌعٌة و النمو االقتصادي المستمر و التزاٌد الدائم لعدد السكان مع أنماط حٌاة‬
‫تستهلك الموارد بكثافة شدٌدة‪ .‬نتٌجة لذلك‪ٌ ،‬تعٌن على صناع القرار القٌام بتحلٌل خٌارات توزٌع و دمج موارد الطاقة المتجددة المتاحة و‬
‫ذلك للوصول إلى حلول ناجعة لهذه التحدٌات‪ .‬الهدف األساسً هو توفٌر خٌارات متنوعة لالستخدام الفعال للطاقة المتجددة فً الحاضر‬
‫والمستقبل القرٌب على حد سواء‪.‬‬
‫‪(wind‬‬
‫تقوم هذه الدراسة التقنٌة‪ ،‬من خالل استخدام برنامج ‪ ، MATLAB‬بتحلٌل إمكانٌة توزٌع منهجً و أمثل لتوربٌنات الرٌاح‬
‫)‪ turbines‬و الخالٌا الشمسٌة (‪ )Photovoltaic panels‬للحد من طلب الحمل على بنٌة تولٌد الطاقة فً ألمانٌا و تقدم حالة دراسة‬
‫مبسطة فً المغرب‪.‬‬
‫تسعى هذه الدراسة إلى تحلٌل نوع السعة المضافة (‪ )Installed capacity‬المطلوبة للتأكد من إسهام الخالٌا الشمسٌة وتوربٌنات الرٌاح‬
‫فً تلبٌة احتٌاجات استهالك الطاقة فً ألمانٌا‪ ،‬و ذلك استنادا إلى أسالٌب التوزٌع المستخدمة فً هذهالدراسة‪ .‬باإلضافة إلى ذلك‪ ،‬تم تقسٌم‬
‫ألمانٌا إلى ست مناطق لمعرفة اآلثار المترتبة على تحسٌن توزٌع الموارد فً المناطق الصغٌرة على النظام برمته‪ ،‬مما ٌسمح لنا أن نأخذ‬
‫النقل ومتطلبات السعة التخزٌنٌة الالزمة لتغطٌة الحمولة القصوى باستمرار فً االعتبار‪.‬‬
‫ٌقلل سٌنارٌو " المناطق الست" المقترح فً هذه الدراسة بسعة اجمالٌة مضافة تبلغ ‪ 238,01‬جٌجاواط‪،‬الطلب على الحمل بنسبة ‪٪69.18‬‬
‫و ٌؤدي بالتالً إلى تغطٌتها بشكل أكبر من ال ‪ ٪ 65‬فً سٌنارٌو ألمانٌا كمنطقة واحدة ‪.‬باإلضافة إلى ذلك‪ ،‬بلغ متوسط تدفق الحمل لسعة‬
‫خطوط النقل متوسط ‪ 1474‬مٌغاواط من خالل شبكة تٌار مباشر )‪ (DC‬بٌن المناطق الست لتحقٌق توازن بٌن فائض و نقص انتاج الطاقة‬
‫كل ساعة‪ ،‬قبل أن ٌتم فً النهاٌة استخدام سعة تخزٌنٌة بمدى ‪ 17‬جٌجاواط وسعة مركبة ‪ 141,737‬تٌراواط‪.‬ساعة ما ٌعمل على تعوٌض‬
‫النقص بٌن اإلنتاج و االستهالك فً كل ساعة وعلى مدار العام‪.‬‬
‫تمت مقارنة نتائج متطلبات السعة التخزٌنٌة و سعة النقل بسٌنارٌوهٌن آخرٌن‪ :‬سٌنارٌو أللمانٌا كمنطقة واحدة وسٌنارٌو مرجعً آخر‪،‬‬
‫وأظهرت النتائج أن السعة التخزٌنٌة القصوى منخفضة فً سٌنارٌو المناطق الست بمقدار ‪ %27‬عن سٌنارٌو ألمانٌا كمنطقة واحدة و‬
‫‪ %12‬عن السٌنارٌو المرجعً‪ ،‬كما وجد أن متطلبات السعة التخزٌنٌة منخفضة بمقدار ‪ %13‬عن سٌنارٌو ألمانٌا كمنطقة واحدة و ‪%2‬‬
‫عن السٌنارٌو المرجعً‪ .‬ما أكد لنا فائدة استخدام سٌنارٌو مناطقً كوسٌلة لتقلٌل متطلبات النقل والتخزٌن تأخذ بعٌن االعتبار النظام ككل‪.‬‬
‫توفر دراسة الحالة فً المغرب تقدٌرات نظرٌة للسعة المضافة المطلوبة لكل من الخالٌا الشمسٌة وطاقة الرٌاح الالزمة لتغطٌة الحمل‬
‫الكهربائً بشكل كامل والتً وجد أنها تساوي ‪ 19520‬مٌجاوات و ‪ 12250‬مٌجاوات للخالٌا الشمسٌة وطاقة الرٌاح على التوالً ‪.‬تمت‬
‫أٌضا دراسة متطلبات التخزٌن لكل حالة و تحلٌل الفروق بٌن خصائص التخزٌن بهذا الحجم لكل مصدر حٌث ظهر أن متطلبات السعة‬
‫التخزٌنٌة للخالٌا الشمسٌة أقل بنسبة ‪ %27‬عن طاقة الرٌاح‪.‬‬
‫كاسل‪ ،‬ألمانٌا‬
‫‪28.02.2012‬‬